Yo u r G u i d e t o

The Wild Jungles Of

Conversion Rate

IN
S.
Optimization
AD
LO
N
W
O
D
SO
W
 Table of Contents

Introduction (p.1)

Chapter I: Some Terminology (p. 1-6

Types of test

A/B testin

IN
Multivariate testin

S.
Split testin

AD
Personalizatio

Types of visual dat

LO
Heatmap

Scroll map N
Session recording
W
Types of conversion
O

Other research methods

Chapter II: Variables to Test (p. 7-10

SO

Layou
W

Navigation men

Headline (H1

Body tex

Call-to-action button

Imager

Video

Social proo

Forms

Chapter III: Tech Stack (p. 11-13

A/B Test Too

Visual insights tool

Landing page builder

CRM

Qualitative research tool

Data enrichment tool

Chapter IV: Laying the Groundwork for Your First Experiment (p. 14-16)
Come up with the experiment idea

Rank your idea

Define your conversion goa

Define succes

Formulate a hypothesis

Chapter V: Ensuring experiment validity (p. 17-19

IN
Choose an audience segmen

S.
Reduce the margin of erro

AD
Ensure that your users are unaware that they’re a part of the experimen

Run your experiment in isolation

LO
Chapter VI: A/B Test Ideas (p. 20-49)

Chapter VII: Things to avoid (p. 50-51 N

Browser notification
W
Newsletter pop-up
O

Auto-play videos with soun

D

Cheap traffic to build up sample siz

SO

Spin the wheel and other gambling-related techniques

Chapter VIII: Analyzing the experiment results (p. 52-53

W

Did you attain an adequate sample size

Have your findings passed the “statistical significance” test

Were there any outliers in your dataset

Was your test running for a sufficient time?

Conclusion (p.54)

Footnotes

Introduction
Are you struggling to convert visitors into leads or paying customers?

Conversion Rate Optimization is a wild wasteland of marketing where every little detail can

have a massive impact on your business.

Throughout this book, we’ll explore the theoretical foundation of CRO and the best practices

IN
for website experiments as well as review a range of tested and effective growth ideas.

S.
This e-book will equip you with all the knowledge you need to double or even triple your

conversion rate.

AD
Whether you're an entrepreneur, marketer, or SEO specialist, your website is the backbone of

LO
your entire digital presence, and CRO isn’t something that you should overlook.

But, enough with the introductions - grab your favorite snack and a notebook, and let’s dive
N
W
right into it!
O

Chapter I: Basic Terminology

D
SO

We'll start the exploration of CRO projects’ wilderness by covering the basic theoretical
W

concepts you'll encounter along the way.

Types of tests
Running CRO experiments can go both ways - you can test small details or go crazy and

revamp your entire page.

Let’s have a quick rundown of the types of tests, and how they suit your needs.

A/B testing

As the name suggests, A/B tests revolve around changing a single variable and comparing it to

the original version. For example, changing the button color from blue to red.

1
 Note: all other elements should stay the same. If you test multiple things at the

same time, you will not tell which variable made a difference.

Multivariate testing

IN
A multivariate test is the same as the A/B test, but this time we’re testing multiple variables at

S.
the same time. For example, if you tweak your CTA color and hero image, you'll get 4 variations

in total

AD
LO
New CTA + Original Imag

Original CTA + Original Imag N

W
Original CTA + New Imag
O

New CTA + New Image

D
SO

Although multivariate tests are efficient, they require a lot more sample data. Since you have

more variations, it will take longer to reach statistically significant results.

W

Split testing

Split testing involves testing 2 different pages. Essentially, the tool will redirect your traffic to

another page with a 50% chance, allowing you to test a homepage vs. a specific landing page.

Personalization

Personalization is a landing page version, tailored to a specific user segment. You can

personalize anything from specific wording to the entire page design. A personalized landing

page is particularly relevant for the following cases

You run a promotional offer that is relevant to a particular sub-audience (s

You run Google Ads and want your copy (title, subtitle) to match the user’s search quer

You need to adjust the messaging for different user segments

2
You can set up a personalized version either through an A/B or split test, depending on the
variables you want to test.

Types of visual data

Apart from your numeric KPIs analysis, visual data might reveal useful user behavior patterns.

IN
Heatmaps

S.
Heatmaps show you the areas of the page that users most frequently hover over with their

AD
mouse. They come in handy when you are evaluating a page layout/ user interface – you get to
see how users navigate your page and the page elements that attract the most attention.

LO
N
W
O
D
SO
W

Source: Hotjar

Scroll maps
As the name suggests, these maps show you how far your users get when scrolling your
landing page. They are useful if you want to evaluate the order of your sections (based on
potential impact) or the layout of your page.

3
Keep in mind that most people won’t scroll all the way down to the footer, so you can use scroll
maps to see how your audience navigates your website.

IN
S.
AD
LO
N
W
O

Source: Microsoft
D
SO

Session recordings

It's an ethically debatable, but insightful technique. You can record the entire browser session
W

of your audience and see all of their actions. How long they fill out the form, how many buttons
they press, how far they scroll etc. 

The upside here is clear – you get to see the entire user journey. 

The downside is that each recording will differ from the rest and watching many of them will
take forever. Besides, visitor recordings are powered by heavy JS, which can notably slow
down your page.

Use this technique to analyze small-scale experiments or for ideation purposes.

4
 IN
S.
AD
Source: PostHog

LO
Types of conversions N
Conversions are usually grouped into 2 types
W
Hard conversions - these are purchases, booked calls, or any other actions that directly
O

contribute to your marketing success.

D
SO

Soft conversions - these conversions revolve around user engagement and intent, such as
starting the checkout, looking at the pricing page, adding items to the cart, or downloading a
W

free asset in exchange for an email.

You can (and should) optimize both conversion groups, but focus your efforts on one depending
on your niche.

Other research methods

Website surveys
Surprise, surprise - you can get real user feedback without advanced analytics or software.
Sometimes the best way to discover your audience's thoughts is to ask them. A lot of websites
have a small feedback form somewhere in the corner of the screen, while others display the
form after a certain action is taken (e.g. after a person uses the live chat or submits a request).

5
In any case, make the form as simple and intuitive as you can. Ideally, you’d ask 1 or 2 questions

at most. Use simple words, emojis, and anything else that can convince a user right away that

they won’t be bored by helping you out.

Also, keep open questions at a bare minimum or avoid them entirely since they are naturally

perceived as time-consuming.

Keep in mind that only a fraction of your audience will actually bother to give you feedback, and

IN
your survey responses will likely be heavily skewed toward negatives. Realistically, most people

S.
with a positive impression of your website will ignore your feedback section.

AD
Interviews & focus groups

LO
If you want in-depth feedback from your website, interviews and focus groups are a great way

to get it.

N
W
Look for participants among your customers or industry experts and get in touch with them.

If you want more people to help you out, try offering incentives like gift cards or discounts on
D

your products or services.

SO
W

6
Chapter II: Variables to Test
In theory, you could test any element of your website. Yet, most landing pages share the same
anatomy, so let’s look at the page elements that should be tested regularly.

Layout

IN
When browsing the web, users tend to skim through the website for just 7 seconds on average.

S.
If they find something too difficult to comprehend - they will leave your website right away.
Thus, it’s your job to make sure that your offering is present in a way that’s easy to digest.

AD
The first place to start would be the layout of your hero section. Is it a pyramid, a single column,

LO
or 2 columns? There’s no right or wrong here because every case is different, but it’s something
that you should test at the dawn of your journey.
N
W
Navigation menu
O
D

As you probably know, your landing pages shouldn’t even have a navigation menu because
SO

landing pages are designed to push prospects to take one specific action. Any extra links or
navigation items cause a distraction toward conversion. 

Nevertheless, you can still use your header section for growth experiments, or for example, by
adding a secondary CTA, social proof, or supporting text. In this case, it’s best to set the header
position to “fixed” (this way, the section will always be at the top of the page, even if a user
scrolls all the way to the footer).

Headline (H1)
It's the first thing a user is going to read. Your headline must be powerful and captivating
enough to entice the person to read further. Therefore, you could test several headlines to
determine which one wins the minds (or hearts) of your target audience.

7
Take a look a few general best practices
Communicate one clear benefit - don’t try to capture all your value propositions in one
sentence, and focus on what you do bes
Avoid jargon or buzzwords - stick to day-to-day language that everybody understand
Keep your headline short - there’s no right and wrong here, but try to fit your headline into
2 lines of text

IN
S.
Body text

AD
In terms of content, headlines usually make the biggest difference. However, people will still
skim through your body text, so your supporting copy should make it more enticing to convert

LO
as well.
N
W
Call-to-action buttons
O

Unsurprisingly, your call-to-action is paramount for your conversion rate. You want your CTA
D

button to be both attention-grabbing and simple. Besides experimenting with button text, you
SO

could also test different colors and shapes of the button (but try to stay true to your brand
W

identity and don’t go too crazy). Here are some tips

Consistent copy - if you have multiple instances of the same call-to-action, use the same
text inside your butto
Visual weight - make your button stand out from other elements by using an eye-catching
accent color and additional visual cue
Simple font - don’t go too fancy with fonts, and pick any sans-serif typeface that’s easy to
read
Generous spacing around the button - give your CTA buttons plenty of breathing space,
and don’t position them too close to other page element
CTA above the fold - always feature a CTA button in your hero section, since it will yield the
majority of your CTR in most case
One CTA per section - don’t make your call-to-actions compete for attention, and make sure
that only a single button is visible in the user’s viewport

8
Imagery

Visual content inevitably attracts more attention than text. Unfortunately, more often than not,
marketers use stock to search for generic photos or vectors that are somewhat connected to
the product they sell. Most of these visuals weren't designed with conversions in mind in the
first place. They are frequently reused throughout the internet and can create an impression of
a template (read as "cheap and lazy").

IN
It would be best to convey the benefits of your product or service through an original image. It

S.
can include emotional benefits (e.g showcasing how people interact with your brand) or

AD
intellectual benefits (e.g adding your key value props to the image).

When it comes to testing, instead of trying to change all of your website imagery at once, test

LO
different visuals on your hero section. Once you get the results - apply the same approach for
all other sections, if needed.
N
W
O

Videos
D

Remember, we mentioned that images are often treated as space-fillers rather than conversion
SO

triggers? Unsurprisingly, the same goes for videos.

A lot of companies order explainer videos from Fiverr, where they’re made with templates in
tools like Powtoon, and… it usually looks terrible.

If you have a product to showcase - don’t cheap out and invest in high-quality video
production. 

Moreover, some tools (such as Vidyard or Wistia) allow you to add links and CTAs to your
videos, which is definitely worth exploring if you plan to use videos frequently in your marketing
funnel.

Social proof

We all know that social proof has a huge effect on consumer perception. However, proof points
can take many forms (testimonials, badges, star ratings, reviews, NPS score, etc.), and it’s
something worth testing.

9
Forms

Whether you’re offering a demo or a free lead magnet, you will likely have a form on your page.

Therefore, you can test the number of form fields, their text, and the overall look of the form to

see what converts best.

General best practices

IN
S.
Stick to a one-column layout - it’s easier on the eyes and better optimized for mobile out of

AD
the bo

Indicate optional fields - if your form features any optional information, clearly mark it by

LO
adding “(optional)” at the end of your field text. Asterisks can be easily overlooked
N
Use placeholders - add a clear placeholder text to each applicable form fiel
W
Use “free text” fields as little as possible - instead of asking open questions, give your
O

users a pool of options to select. You can do it by using dropdowns or radio buttons
D

Use the invisible captcha - most websites use Google’s reCAPTCHA service to detect bot
SO

traffic, and that’s okay. Yet, searching for matching images is not a fun process, and a lot of
W

people will quit the form. On the other hand, an invisible captcha runs in the background

and doesn’t complicate the journey of your audience.

10
Chapter III: Tech Stack
In this chapter, we’ll cover the tools you may need on your journey. Some of them are clearly a
“must” while others are nice-to-have, depending on your experiment scope. Dedicate some
time to comparing all the options because updating your tech stack won’t be an easy task in the
future.

IN
A/B Test Tool (must have)

S.
AD
As you might’ve guessed, you’ll need a tool to build website tests. If you’re a developer - you
could launch your experiments with custom code, but that’s tedious and likely not worth it.

LO
Google Optimize used to be a very decent free tool, but since Google is sunsetting it, we’ll be
looking at other options. Most of them offer a trial or a free version with limited features. 

N
W
Here's a list that features the most popular A/B testing software. Personally, I use VWO, but it’s
O

not the cheapest tool around.

D
SO

List of tools: VWO, Optimizely, A/B Tasty, Crazy Egg, GrowthBook

W

When deciding on an application for your experiments, ask the following questions
Does it feature a simple visual editor
How many metrics/ events/ conditions can you track
Do you have enough filters to target specific audience segments
Does it offer the integrations you need (e.g. Google Analytics, Hubspot)?

Visual insights tool (nice to have)

Apart from analyzing hard figures, you might want to look at all kinds of visual data to analyze
or ideate website experiments.

This data includes but isn’t limited to heatmaps, scroll maps, and visitor recordings.

Some software, like VWO, has these features built-in, while other tools don’t. Therefore, you’ll
likely need a dedicated tool for that.

11
Visual insights tool (nice to have)

One of the most popular tools around is Hotjar . The user interface is really simple to navigate,
and the setup can be done in 5 minutes. However, their free plan is extremely limited, so you’ll
likely have to open up your wallet.

There’s a free alternative to Hotjar called Microsoft Clarity . It’s a great option if you want to
save money, as it’s packed with more or less the same features. It lacks some marketing

IN
integrations, and the UI could be better, but again, it’s a no-brainer if you’re on a budget.

S.
AD
Landing page builder (nice to have)

Once again, this is optional, but if you build a lot of landing pages, a dedicated tool could help

LO
to edit and test your designs with ease.

N
There are a ton of great page builders out there, but most marketers go with Instapage ,
W
Unbounce, or LeadPages.

I prefer Instapage, but all of the mentioned tools have an easy-to-use interface, useful
D

integrations, and testing features.

SO
W

CRM (optional)

You don’t need a CRM to run your website experiments. But if you use one - you’ll have richer
data.

Most optimization tools are tailored toward analyzing traffic but not lead quality or deal size,
which is particularly relevant for B2B firms.

For instance, VWO integrates with Hubspot , meaning every time a contact is added to the CRM
- it’s stored with a certain identifier. Thus, you could see if your experiment made an impact on
metrics that you wouldn’t measure otherwise.

Popular CRMs such as Salesforce or Hubspot are usually well-integrated, but they will cost you
a fortune. If you go with a less-known CRM, you might have to set up the integration yourself
via API or Webhooks. 

Do a cost-benefit analysis and decide what works best for you (if you need this functionality at
all).

12
Qualitative research tool (optional)

Finally, launching a website experiment is not always feasible or desirable.

For instance, if you want to obtain general opinions on your design and value proposition

(outside of conversions), you would likely opt for qualitative research methods, such as

interviewing or focus groups.

Additionally, you probably won’t have enough traffic and conversion data to get significant

IN
results for every experiment. If that’s the case, you might want to look at qualitative research

S.
tools like Wynter.

AD
Tools like this help you find people that match your ideal customer profile and are willing to

LO
share their thoughts. Needless to say, it’s always paid if you involve a middleman.

Data enrichment tool (optional)

N
W
The subject of data enrichment and its use in CRO experiments likely deserves a separate
O

book, so I’ll keep the following section short.

D
SO

There are a variety of tools on the market with immense datasets that help you to identify user

information based on a single website visit. This is particularly helpful for B2B markets, where a
W

prospect usually undergoes a lengthy sales cycle.

An example of such a tool is Clearbit . Apart from retaining the information in your CRM, you can

connect the dataset to your tech stack and personalize user experiences in real-time (e.g.

specific form fields, dynamic landing pages, personalized call-to-action buttons, etc.)

Alternative tools: Albacross , Leadfeeder , Leadinfo

A few notes

Data enrichment tools tend to be very expensive, so it’s usually not something

that small businesses can invest i

Do your research and make sure that the tool of your choice complies with all

local regulations (especially GDPR and CPDA)

13
Chapter IV: Laying the Groundwork for
Your First Experiment
We’re doing great. By now you should have a pretty good notion of basic theory and a range of
marketing tools that can help you along the way. So, it’s time to prepare the foundation for your
first CRO experiment.

IN
Come up with the experiment ideas

S.
The first (and often the hardest) step is coming up with a solid idea for an A/B test. Luckily, it’s

AD
not that difficult to get inspired. Here are a few effective ideation techniques that you can try in

LO
your workflow
This book contains 42 effective CRO ideas. Even though it's biased, it'll do for a good star
N
W
Take a look at your competitors’ websites. Get inspired by their design, layout, or
O

messaging, and test this idea on your business website

D

If none of your competitors inspire you, peek into other industries. Start with well-known
SO

marketing websites (e.g. Intercom), as some of them are already well-optimized for
conversions
W

Finally, you can do online research or ask chatGPT to generate some ideas for you. You’ll
see a lot of the things that you came across before, but you might also find something new.

Rank your ideas

Say you’ve got a list of solid ideas for growth experiments. Now it’s up to you to prioritize and
rank them. One of the most common frameworks used by growth specialists is called “ICE”
(unrelated to Vanilla Ice).
It stands for
Impact - how much value do you expect to gain from this experimen
Confidence - how confident are you that the experiment will be a succes
Ease - how much time and resources it will take to launch to experiment

14
All you have to do is add a score to your ideas for each point (from 1 to 5) and divide by 3 to get

the mean average. Subsequently, you start with the ideas that have the highest score, and,

thus, the highest potential.

Note: these frameworks are particularly helpful when you’re working with a team. If you’re a

solo entrepreneur, your best bet is trusting your judgment.

IN
Define your conversion goal

S.
What result are you trying to achieve? More downloads, higher time spent on a page, lower

AD
bounce rate? Make sure that each experiment that you run has a clear and measurable goal

LO
behind it.

While optimizing for a specific conversion, don’t forget to track any important secondary metrics

in the background.

N
W
It's vital because, for instance, your a/b test might do wonders for your clickthrough rate, but
O

your sign-up rate might actually decrease. Hence, you shouldn't lose track of post-click
D
SO

conversions when optimizing for button clicks.

W

Define success

Now that you have a goal in mind let’s make it measurable. First of all, take a look at your

current data and try to make a well-educated assumption. 

What is your current conversion rate, and what would be the next realistic milestone? Let’s say

that your current conversion rate is 1%. Going up to 1.5% means a 150% increase. Could this be

achieved by a single experiment? Possibly, but try not to be too ambitious - after all, CRO is a

continuous process.

If you have no available data or can't determine what’s a success and what’s a failure - compare

your performance with industry standards.

Formulate a hypothesis

You found an idea you want to test, a metric that should be affected, and some benchmarks to

judge. Now it’s time to bring this all together and formulate a hypothesis.

15
An example: I think that increasing the size of the CTA button in our hero section will positively
impact the sign-up rate. I expect to see a difference of about 15%.

Unlike the idea you have in mind, a hypothesis is a specific and documented prediction that you
want to test.

IN
As a side note, if you’re testing multiple variables simultaneously, it helps to draw
a conceptual model to capture all the possible relationships between those

S.
variables.

AD
LO
N
W
O
D
SO
W

16
Chapter V: Ensuring experiment validity
Before you go ahead and launch your first experiment, be aware that there will be dangers in
the CRO jungles that a lot of marketers don’t account for.

Choose an audience segment

IN
Start by deciding what segment of your audience you should focus on. Some experiments are

S.
generic in nature, and you can run them on all traffic, while others are more nuanced and

AD
require a specific user group.

Let’s look at a few segmentation options that you can consider

LO
Specific ad channels (Facebook, Google, LinkedIn, etc.
Referral source
N
W
CRM status (lead, client
O

ICP status (fits or doesn’t fit ideal customer profile

D

Marketing funnel stage (BOFU, MOFU, TOFU

SO

Specific locations (geo

User activity (page views, completed events, downloads, etc.
W

Device (desktop, mobile, tablet)

Reduce the margin of error

Perfect research doesn’t exist, and there will always be room for error. However, you can try to
minimize your margin error by doing the following:

Testing elements that are too minor or have too many variations

Be pragmatic when it comes to implementation. Make sure that you’re testing a variable that
can make an impact, not something overly minor.  

If the testing variable is too easy to overlook - it increases the odds that any observed
difference might be random, and not directly related to your experiment.

17
Similarly, try not to test too many variants that are very similar to each other. (for example, very

similar shades of the same color)

Having too many variations can lead you to false positive results, commonly called a “Type I

error” in statistics).

Always split traffic evenly

IN
This point is fairly simple, try to keep the samples of your A and B variations even in terms of

S.
size. Most tools will allow you to control how you want to split traffic to your variation (e.g.,

AD
50/50, 80/20, 60/40, etc.)

There are very few use cases where you’d need to adjust these proportions, so keep it at

LO
50/50, which is usually the default option.
N
W
Exclude internal traffic from experiment targeting

You probably have a Google Analytics filter that excludes internal traffic based on a list of IP
D

addresses, so you can simply reuse it.

SO

Make sure that your conversion event triggers once per session

Most people track their purchases or sign-ups via a thank-you page or form submission.

Let’s say you converted and went on a thank-you page and then reload it - this may count as 2

conversions instead of 1.

Alternatively, there are bots (or bored people) that just submit the form to spam

you, so you’ll see multiple form submissions that might come from the same

source.

If you are optimizing for a macro-conversion, it’s best to let your conversion tag/

event/ tracking pixel fire only once per session to eliminate some of the “junk”.

18
 Never change your variants or settings in the middle of the test

If your test is functioning properly - don’t make any changes to it, as they might impact the

results and lead you to a wrongful conclusion. If you’ve missed something - you can either let

the test run or pause it completely and re-launch it at some point in the future.

Ensure that your users are unaware that they’re a part of

IN
the experiment

S.
You load a webpage and within a second or two, you notice that the CTA text or color has

AD
changed. Could it be a creative decision? Probably, but most likely, it’s a marketer who didn’t

LO
check their setup properly.

Sometimes, your tool or code snippet can load asynchronously (after the main content has
N
loaded), which is bad for the user experience and your test. If users see an element that has
W
been changed upon page load - it might affect their perception, either positively or negatively.
O

It means your experiment’s outcome will not be valid either way, as you’re introducing an
D
SO

element of surprise that could make all the difference. Simply put, make sure that variation A

or B gets pre-selected before the page load.

Similarly, ensure that there are no caching issues. If a user returns to your website an hour

later - they should still see the initial variant they had before. Showing multiple variations of the

same page on a user level can also affect the validity of your results.

Run your experiment in isolation

As mentioned above, many factors impacting conversion rates can be outside your control. It

can be weather, time of day, or literally anything else.

Nonetheless, you can take measures to ensure that your experiment runs in complete isolation

from other internal factors. For instance, you don’t want to test your pricing page during a Black

Friday sale because it can substantially skew your results.

Similarly, you don't want to run multiple tests on the same page at once, as they might

overlap with each other and distort your conclusions.

19
Chapter VI: A/B Test Ideas
Hurray, now you’ve got everything it takes to launch conversion experiments. This chapter
reveals 42 time-tested and effective ideas that you could borrow for your next CRO project.

1 Make the perceived cost seem lower by reframing your

pricing

IN
S.
Category: pricing Difficulty: easy Impact: medium

AD
If you’re offering a product that most people/ companies can easily afford - make it your selling
point.

LO
There was a study done by Carnegie Mellon University where researchers compared the
N
performance of two banners. In one case, there was “a $5 fee”, and in another, it was a “small
W
$5 fee”. Even this slight change in the text made a difference in terms of conversions.

You don't have to make it obvious that your price is low - simply reframe it instead. For instance,
D

your product is worth just 2 cups of coffee. Giving up $5 is perceived as a bigger investment,
SO

compared to giving up 2 cups of coffee.

Similarly, you can frame the pricing of your monthly plans in “per day” cost, making the
perceived pricing seem lower than it actually is.

2 Add visual cues to draw attention to your CTA button

Category: layout Difficulty: medium Impact: medium

There's a principle called "leading lines" in photography and art, which implies framing the
composition in a way that naturally leads the human eye toward the point of interest. 

We can re-use this technique to optimize your landing pages and draw more attention to our
CTAs.

20
Let’s look at a few visual cues that you can try

Your image can include a person gazing/ pointing at your CTA butto

You can add an arrow pointing to your button. It's bold and straightforward, but it works well,

as long as it fits your design

Your image can include an object that faces the button (thus creating a leading line).

IN
S.
AD
LO
N
W
O
D
SO

Source: Nudgify
W

3 Highlight the new price against the old price

Category: pricing Difficulty: easy Impact: high

A time-proven hack that never fails to work. Due to anchoring, people often base their

perception of the product price on the first piece of information available to them.

To put it into perspective, you can present your original prices on your pricing page, cross them

out, and put the updated prices next to them.

What will happen is people will perceive your new price as “less expensive”.

21
 IN
S.
AD
LO
N
Source: Zalando
W
O

4 Frame your copy around loss aversion

D
SO

Category: content Difficulty: medium Impact: medium

As behavioral economics taught us, people hate losing something more than they like winning
W

the same benefit.

Thus, instead of emphasizing that your users will save X hours or X amount of money with your
product, you can instead stress that they will avoid losing the same time or money.

5 Use social proof notifications

Category: emotional triggers Difficulty: medium Impact: medium

You can display widget notifications to showcase dynamic proof points to your visitors. 

Let’s look at a few examples

“300 people have signed up for our webinar in the last 7 days
”50 growth marketers bought this course yesterday
“70 frequent flyers are looking at these tickets now”

22
There are plenty of tools on the market that can record this data and display it on your website

(e.g. ProveSource , Nudgify , WiserNotify ). Needless to say, you can also make it work without

any 3rd party tools, but you’ll probably need to be comfortable with APIs and JS.

An extra benefit of using a dedicated tool comes from the fact that it creates an impression that

the data is verified. Even though you can always game the system and fake the data, that’s not

what I’d advise you to do.

IN
S.
AD
Some extra tips on this one

Try not to get too personal. (e.g. display only data of groups and not

LO
individuals). Something like “John Smith from the UK” is a privacy violation

Display one notification per page, and only once. If it doesn’t work - so be it,
N
W
but you don’t want to annoy your visitors.
O
D
SO
W

Source: Nudgify

23
6 Use the decoy effect when coming up with pricing
packages
Category: pricing Difficulty: medium Impact: medium

There’s a term from behavioral economics called the “decoy effect”. It entails that a certain
option will seem more favorable if presented next to the other option, which is clearly inferior.

IN
For example, if you make your premium package seem expensive - with this context, your

S.
medium package will seem more favorable. The same logic applies the other way around - you
can make your basic package very limited to make your desired offer stand out.

AD
Long story short, if you have pricing plans, always put your best offer in the middle, and reframe

LO
packages so that one offer is superior to the rest.
N
W
O
D
SO
W

Source: Squarespace

7 Use the default effect for upselling your service

Category: pricing Difficulty: easy Impact: medium

This is closely related to the point above. People tend to go for the default option when given a
choice (this is called the default effect). If you have a desired offer in mind (say, extra services
or upsells) - use the default effect to your advantage.

24
Let’s look at a familiar real-life example – remember the last time you bought an airplane ticket?
Most airlines have a variety of extra services (e.g., meal on board, no waiting time, extra
luggage, and so on) – and if you go and check out their purchase form, most of them will be
checked by default. 

You might be one of those guys who unchecks everything, but a lot of people don’t. Use this
technique carefully because you don’t want to be too pushy.

IN
8 Add complementary text under your CTA button to

S.
reduce friction

AD
Category: layout Difficulty: easy Impact: high

LO
When you prompt your visitors to convert - what is the #1 thing that can hold them back?

N
Hesitance to provide credit card details, fear of checkout taking too long, or a long-term
W
commitment are some of the doubts your prospects might face.

Add an extra line of text under your main CTA that can reduce this friction. A few examples
D
SO

If you’re offering a free trial with no credit card required - mention it

If you’re offering a free consultation - indicate there’s no commitment.
W

Note: you don’t necessarily have to put this copy under the button, but it’s usually
preferred because the point of friction is grouped with a point of action.

25
 IN
S.
AD
LO
N
W
Source: ClickUp
O
D

9 Showcase your guarantees above the fold and stress

SO

the absence of risks

W

Category: layout Difficulty: medium Impact: medium

Since we all have a tendency to go for risk-free products, it would be unwise not to capitalize
on this.

If you offer any form of guarantee (e.g., money-back guarantee, satisfaction guarantee, delivery
time guarantee) - always include it above the fold.

All these safety nets serve as powerful conversion triggers - make people notice them right
away.

26
 IN
S.
AD
LO
N
W
O
D
SO

Source: Landingi
W

10 Highlight a trusted party that brought the visitor to

your website
Category: layout Difficulty: medium Impact: medium

If your traffic comes from a review website, marketing partner, or any trusted 3rd party - why not
mention it on your landing page?

You can add a “badge” in your hero section that says something like
In cooperation with XY
Shortlisted by XY
In partnership with XY
Supported by XYZ

This way, the prospects will form a stronger link between your business and the party that
brought them to you.

27
 IN
S.
AD
LO
S o u r c e : Re c r u i t e e

N
W
Note: always ask permission from your partners to reference them or use their
O

logos. Also, don’t use the following technique to reference ad channels

D

(Facebook, Google).
SO
W

11 Every time you run a promotion - put it in the spotlight

Category: sales Difficulty: medium Impact: high

When you’re running a sale - you want everyone to know about it.

Below are a few ways to attract more attention to your promo without modifying the original

content of your home page

Add a notification bar to the top or bottom of your pag

Use your live chat to display a small notification banne

Display a small pop-up with the coupon code on your checkout page

28
 IN
S.
AD
Source: Shopify

LO
12 Present a different homepage experience for logged-in
users or captured leads N
W
Category: personalization Difficulty: hard Impact: high
O

Most websites serve a static homepage to all groups of visitors, and they’re missing out. Let’s
D

say your homepage is meant for people to convert to your free trial, but what about repeat
SO

visitors who have already converted?Similarly, if people are logged in to your product - there’s
W

no point in prompting them to take a free trial, right?

Keeping this in mind, you can design a somewhat unique homepage experience for people that

have already converted. For instance, you can change your CTAs from “take a free trial” to “buy

now” or invite users to check out your blog or any other supporting content.

Most tools allow you to design a dedicated page experience by creating an audience of visitors

based on the criteria you define. Yet, with some elbow grease, you can pull it off yourself using

cookies, Google Tag Manager, and a bit of custom JavaScript.

Note: changing too many layout-related variables with JS can slow down your

website, so make sure to test thoroughly.

29
13 Match the page copy to the search intent of the user
Category: content Difficulty: medium Impact: medium

This tip is for you if you run Google Ads or any other ads based on search intent.

Most of the time, we see companies bidding for hundreds of different keywords and linking
their homepage.

IN
This approach is ancient, and you’ll find that making dedicated landing pages is usually much
more rewarding. However, you can also use the following method for your homepage if you

S.
wish to.

AD
The idea here is simple - take your search term (or its category) and use it in your H1. For

LO
example, let’s say you’re bidding on the following keywords: marketing automation software,
marketing automation tool, or marketing automation system. All you have to do is modify your
N
W
headline to include the keyword that matches the user’s intent.
O
D
SO

If you’re doing it with multiple landing pages - simply change the link in applicable
ads. If you’re doing it for the homepage - you can set up A/B tests/
W

personalizations and use UTM properties (such as utm_term) as targeting criteria.

Source: Zoho

30
14 Add icons to your form fields
Category: layout Difficulty: easy Impact: low

As we previously discussed, visuals attract much more attention than text. Consequently, your
brain is used to seeing the icons within the UI and attaching meaning to them before you even
read the text.

IN
Try adding small icons to your form fields (name, email, company, etc.). You’ll be surprised, but I

S.
tried this once, and the form submission rate went up by 8% - not bad for such a small change.

AD
15 Use a live chat app to display marketing notifications

LO
Category: personalization Difficulty: medium Impact: medium
N
Most businesses use a live chat app like Intercom, Freshchat, or Drift.

W
Yet, the majority display a generic message like “get in touch with us, we’re online, blah blah
O

blah”.

D
SO

You can actually utilize your live chat to display actionable notifications that can push people
W

toward converting. Let’s look at a few examples

Say a prospect has been looking at your page for over 30 seconds without taking action.
Display a message that prompts them to book an introductory call with you
If you’re offering lead magnets such as free e-books or webinars, use your chat to display a
notification on your blog that would prompt people to check it out. It usually attracts more
attention and yields a higher CTR than a static banner
If you’re running a promotion/ sale - you can notify users about it
If an existing customer revisits your website - you can prompt them to leave a review.

As you can see, you can set up many experiments with a simple live chat app.

In terms of automation, Drift offers the most flexibility, but all the options mentioned above will
do the job just fine.

31
 IN
S.
AD
LO
Source: Tidio

N
W
16 Reduce the number of pages in your conversion funnel
O

Category: flow Impact: medium Difficulty: medium

D
SO

Every additional page in the user journey increases the chance of drop-off.

We often see short landing pages with a CTA that leads people to a separate page with a form
W

( just a form). 

You can place the form right at the bottom of the page instead and make your CTA scroll down

the page. You can do it by adding an id to your URL (e.g. website.com#booking-form).

17 Reduce friction in your checkout page

Category: layout Impact: high Difficulty: medium

E-commerce business owners, this one's for you.

People don’t like long checkouts, and one way to reduce your drop-off rate is to display the

checkout stages. Thus, you can design a multi-step checkout process or make a progress bar -

anything that will help the users estimate the amount of time or effort it will take to complete

the checkout.

32
 IN
S.
AD
LO
N
W
O
D
SO

Source: Moz
W

18 Offer split payments or employ a subscription-based

pricing model

Category: pricing Impact: high Difficulty: hard

Companies love subscriptions for multiple reasons. First, people tend to spend more money

when it’s denominated in small amounts or installments, as opposed to paying a large lump

sum. Second, let’s be honest, companies like Netflix profit greatly from consumers who simply

forgot to unsubscribe.

If you sell a product rather than a service, definitely offer the option to pay in monthly

installments. It can increase your sales by a huge margin.

33
 IN
S.
AD
LO
N
W
O
D
SO

Source: Ben
W

19 Use a psychological pricing strategy

Category: pricing Impact: low Difficulty: easy

These pricing tactics are well known among marketers and salespeople, and at this point, they
have been used for generations.

Here are a few examples that you’ve seen on numerous occasions

Remove one cent from your final price (e.g. $3.99 instead of $4). People in most parts of the
world read left to right, and the first figure will anchor in their minds
Remove the zeros (e.g. $3 instead of $3.00). The less space your price tag takes - the lower
the price seems
Adjusting the font size of prices to make the perceived costs seem smaller. (e.g. smaller font
size for regular prices, larger font size for deals)

34
 Note: the strategy above is only applicable to common products where the price
is often the deciding factor. Do not attempt to use it for premium goods because,
in this case, the items should feel expensive.

Note 2: Do not use these strategies for promotions. In this case, it’s better to stick
to exact numbers or percentages.

IN
S.
20 Call out your audience in your ad headline

AD
Category: content Impact: medium Difficulty: easy

LO
Whenever you run a marketing campaign with broad targeting - you can attract the right
N
audience by calling them out right in the headline.

W
Examples
O

Start-up founders, this tip is for yo

D

B2B Marketers, this tool is going to supercharge your workflow

SO

People who don’t associate themselves with the following job titles or characteristics will likely
W

ignore your ads. However, most clicks from your ads will come from people who match your
ICP.

21 Experiment with personalized offers

Category: e-commerce Impact: high Difficulty: hard

You’ve noticed that most e-commerce websites rely on personalization to close more sales or
generate more upsells.

This is an interesting testing ground, as there are multiple types of personal recommendations
Recommendations based on pages you viewed or searches that you’ve completed (e.g.,
user activity
Recommendations based on what other people have bought (e.g., group activity
Recommendations based on relationships between the products that you’ve looked at (e.g.,
product-based advice)
35
Thus if you’re running a serious e-commerce business - experiment with different personalized
offers to determine what works best.

IN
S.
AD
LO
N
Source: Amazon
W
O

22 Reduce the number of form fields through data

D

enrichment
SO

Category: layout Impact: high Difficulty: hard

W

The longer your form is - the fewer people will fill it out.

Whenever you want to add an extra field to your form, ask yourself if this is really necessary.

For instance, a lot of B2B firms will often ask for a company name, company size, and industry
in their lead gen forms.

In reality, all you need to ask is the company name. Any data enrichment tool or even a
Linkedin scraper (e.g., Albacross, Clearbit) can get you all the company information you need.

In fact, if your forms only accept business emails, you might not even need to ask for a
company name - in most cases, you can extract it from the email address. 

I’ll leave the implementation and technicalities up to you to decide, but the point I want to make
is - do your research and make the life of your prospects easier.

36
 IN
S.
AD
LO
N
Source: Clearbit
W
O

23 Add figures and emotional hooks to your headlines

D
SO

Category: content Impact: medium Difficulty: easy

This tip is particularly helpful for writing headlines for your ads and emails.

Make your reader feel like they’re about to know something special. For example

Only 5% know about this..

This is what 90% of marketers get wron

9 out of 10 recruiters wish they'd known it earlier

This copy immediately arouses interest and makes people feel like they’re about to get an

advantage over their peers.

24 Create a sense of urgency or scarcity

Category: sales Impact: medium Difficulty: easy

24 Create a sense of urgency or scarcity (tag: sales) (tag: medium impact) (difficulty: easy) 

At this point, it's conventional wisdom that emotional triggers help to generate sales.

37
Here are a few ways that you can use FOMO (fear of missing out) to your advantage
Every time you run a promo, always indicate its time frame in all relevant content (pricing
page, email, ad
Display the number of products left in stock when it’s below a certain threshold (applicable
for e-commerce
Add a timer that counts the time left until the promo ends and display it on your landing

IN
pages or even checkout page

S.
AD
LO
N
W
O
D
SO

Source: Microsoft Xbox Game Store

W

25 Reduce your page load time

Category: technical Impact: high Difficulty: medium

One of the most important CRO initiatives you can ever take is optimizing your website for fast
loading times. Every extra second of page load forces a lot of people to leave early and kills
your conversion potential.

The website optimization process deserves a guide on its own, so let’s keep it brief
Monitor your website regularly using tools like Pingdom or PageSpeed Insights by Googl
Choose a reliable and fast hosting provide
Resize and compress your imager
Load external scripts asynchronously (after the page has loaded)

38
 IN
S.
AD
Source: PageSpeed Insights

LO
26 Use additional cues to highlight your optimal pricing
N
plan
W
O

Category: pricing Impact: high Difficulty: easy

D

Visual cues can help to make your #1 package stand out even more. You can add extra visual
SO

weight to it by switching to a high-contrast color or adding a border or shadow to it.

Additionally, you can add a tooltip that says “top choice”, “most popular”, or “best offer” to

attract even more attention.

Source: SendGrid

39
27 Add a bar with key benefits at the top of your landing
page

Category: layout Impact: medium Difficulty: easy

E-commerce websites are structurally heavy by nature, and your objective is to retain the user’s

attention for as long as possible.

IN
A space-efficient way of highlighting your main selling points is to add a bar above your top

S.
menu.

AD
You can include a few UVPs, such as “free delivery”, “30-day warranty”, and “money-back

guarantee”.

LO
N
W
O
D
SO
W

Source: Zalando

28 Push people to take one specific action

Category: layout Impact: high Difficulty: easy

In most cases, I’d strongly recommend including a single CTA across your landing page.

However, having a primary and a secondary CTA is also common.

If you’ve given it some thought and decided that you have to give your users 2 options, then

follow one of the following practices

Present both buttons but make one CTA visually dominant. You can use color, contrast,

and even animation to make one button more eye-catching than the other

Add your secondary CTA to the navigation menu. In this case, most of your users will

convert for the primary CTA, but the secondary CTA will still be visible.

40
 A common example is adding a trial CTA on all sections of the page while keeping

the option to “buy now” in the top menu.

29 Segmented remarketing

IN
Category: ads Impact: high Difficulty: medium

S.
Here’s a simple tip to make your remarketing campaigns much more effective.

AD
More often than not, companies will run the same retargeting ads on all website visitors.

LO
Refer to our talk about segmentation, and apply it to your remarketing campaigns. 

First off, run specific ads for specific segments - it will pay off. 

N
Second, you’ve probably seen a lot of ads that prompt you to sign up AFTER you have just
W
signed up. In the case of small businesses, this simply confuses your users, but for high-traffic
O

websites, it also costs a fortune to retarget converts with deliberately irrelevant ads.

Finally, if your ads target specific user groups - it also makes it easier to identify certain
SO

patterns and tweak your ads accordingly (e.g. messaging, imagery, key value proposition, etc.)
W

30 Make your CTA stand out by giving it a unique color

Category: design Impact: medium Difficulty: easy

Elements that look untypical to the rest of your design will inevitably stand out. This principle is

called Von Restorff Effect and is widely applied in modern web design.

Let’s look at a practical example.

Most websites use a single accent color for all important elements (links, highlights, buttons,

etc.), and this is not a bad thing.

However, this also means that all of your accent points share a common pattern and don’t stand

out as much as they could.

You can try to give your button a different color - the one that complements your color scheme

but is not used anywhere else on the page.

Apply this experiment with caution and consult a designer to make sure you stay on brand.

41
31 Start your forms with less personal questions and
present one form field at a time
Category: layout Impact: medium Difficulty: medium

First of all, mandatory sharing of personal information right off the bat might drive a lot of
people away, whereas if you start by asking more general questions - people will be more

IN
inclined to start filling out the form.

Second, every form is a micro-commitment. Presenting a single field at a time can make the

S.
overall experience feel less demanding.

AD
LO
Let’s consider a practical example. Let’s say that you have the following form fields: full name,
email, company industry, and job title. Start your form with less personal questions, such as job
N
W
title or industry, and present a single field at a time to make the form feel like a quiz.
O

32 Show the most promising CTA depending on the ICP

D

factor
SO
W

Category: design Impact: medium Difficulty: hard

This experiment idea is particularly applicable to B2B SaaS companies.

Most software companies have 2 conversion funnels, one starting with a free trial and another
with a product demo or sales call.

Now, more often than not, B2B firms would prefer to hop on a call with leads that match their
ideal customer profile and let their sales teams do their magic.

Thus, you could dynamically change your CTA based on potential customer fit and only prompt
more promising leads to book a demo.

Your existing tech stack most likely tracks visitor’s location - that’d be a decent start.

However, if you own a data enrichment tool that can supply company data in real-time, you
could even set conditions based on more advanced criteria such as industry or company size.

42
 Note: you don’t need to offer demos or trials to implement this experiment idea.

The idea is worth a shot if you

Offer 2 distinct conversion funnel

Have a clear ICP (ideal customer profile) in min

IN
Attract a broad pool of website visitors that transcend different geolocations, industries, job

S.
titles, or any other characteristics relevant to your customer fit assessment

AD
33 Experiment with discount framing

LO
Category: pricing Impact: medium NDifficulty: easy
W
Similarly to pricing, your way of framing a discount can significantly affect your sales.

For example, you can try the following

D

If your pricing is relatively small, always display promos in percentages. For example, your
SO

product costs $100. In this case, a 25% discount sounds superior to a $25 discount
W

If your pricing is high, switch to a precise amount. For instance, your product costs $800. In

this case, a $100 discount sounds more enticing than a 12.5% discount

If you offer subscriptions - frame your discounts around free usage instead of price

reduction. For example: “2 Free months with annual payment” sounds more attractive than

“17% discount on annual plans”.

Note: Just remember to stay consistent. Don’t try to combine percentages and

precise amounts in the same promotion, as this will only lead to confusion.

43
34 Offer a unique payment method
Category: pricing Impact: medium Difficulty: hard

You’ve probably come across multiple cases where a purchase value is denominated in some
form of credits or points rather than dollars or euros. This method is frequently used by video
game publishers, mobile apps, and e-commerce shops. 

IN
The underlying logic is that using an arbitrary currency for transactions takes away the pain of

S.
payment. Thus, your products would be sold for “credits”, whereas credits can be purchased
with fiat money.

AD
If you’re operating within B2C, you can certainly try to offer & emphasize your custom payment

LO
method, but be aware - things can go very wrong. It might confuse your audience and create
an impression that your business sells “sketchy” products or services.

N
W
Treat this idea with caution and collect a lot of data before deciding to roll it out.
O
D
SO
W

Source: Rocket League, Epic Games Store

44
35 Add a strikethrough effect to missing features in your
basic pricing plan
Category: pricing Impact: low Difficulty: easy

If you’re offering multiple pricing plans - it’s in your best interest to convey the idea that your
top-tier plans offer much more value than a basic plan.  

IN
It can be achieved by optimizing your plans to trigger FOMO (fear of missing out) and loss

S.
aversion. One way to do it is to include all the missing features in your basic plan and add a
strikethrough effect to the text.

AD
This way, all the absent features are still visible, and it’s clear what your prospects can

LO
potentially miss out on.
N
36 Make use of the “serial position effect” to reorder page
W
elements
O
D

Category: layout Impact: low Difficulty: easy

SO

The term “serial position effect” refers to our tendency to remember the first and last items in a
W

series better than the ones in the middle. 

You’ve probably noticed that this effect is prevalent in our day-to-day life (e.g., remembering
the beginning and the end of the movie more than the middle part).  

Unsurprisingly, we can make use of it to optimize your web experience. Here are 2 quick
examples

Rearranging the links in your navigation menu - most important links first and last, less
important ones in the middl
Rearranging your value propositions within your sections, putting the least important
points in the middle

45
 Note: To observe any difference, make sure that a person will see all series items
at once. For instance, you can’t reliably rearrange your page sections because
you don’t know when a user will decide to stop scrolling. The navigation menu,
however, is a different story because the entire collection of links is visible at
once.

IN
S.
37 Use first-person speech in your call-to-action

AD
Category: content Impact: low Difficulty: easy

LO
First-person speech implies a sense of ownership and control and is often tied to a higher click-
through rate. Some examples
N
W
“Start my free trial” instead of “Start your free trial
O

“Grab my free copy” instead of “Grab your free copy”

D
SO

38 Reward your audience with tangible benefits

W

Category: emotional triggers Impact: high Difficulty: medium

By definition, tangible benefits have a clear value and are easy to quantify.

Let’s look at 2 examples

If you offer an online course, offer a certificate of completion. Not only this benefit
stimulates people to finish the course, but it offers a tangible benefit - it’s something they
can share on their Linkedin, for instance
If you offer a loyalty program - instead of offering a percentage-based discount for the next
purchase, give away a voucher with a fixed discount (e.g., $10).

46
 Source: Coursera

IN
39 Personalize your checkout page

S.
Category: e-commerce Impact: low Difficulty: medium

AD
It's not a secret that people perceive things related to them more positively than generic

LO
experiences.

N
Your checkout page is not an exception to this principle, and you can certainly make it feel
W
personal.

Here are a few methods that you can try

D

Look for placeholders in your checkout form and replace them. For instance, instead of
SO

generic words like “order”, try using more specific words, such as item names or category
names. (e.g., “your jeans can be delivered tomorrow” instead of “your order can be
W

delivered tomorrow”)

Leverage your existing customer information. If a customer makes a repeat purchase, insert
some information you know about them in your copy (start with a first name). For example, "This
jacket will look good on you, Bill" is much nicer to read than "This item has been added to the
cart."

40 Emphasize user autonomy in the decision-making

process
Category: content Impact: low Difficulty: easy

People try to resist if they feel they are being persuaded and their autonomy is restricted.

A good way to overcome this is by reminding users of their freedom to make the decision that
suits them.

47
For example, say you have the following line of a copy somewhere on your page - “we are

confident that we have the best offer on the market”. You can add an extra line of text

underneath saying “you are free to explore other options”. 

It sounds counterintuitive, but this gentle reminder contradicts the user’s expectations (of being

persuaded further) and may come across as more transparent.

IN
S.
This technique applies to short copy, such as a form headline or a call-to-action

AD
button.

LO
41 Offer easy sign-up methods N
W
Category: layout Impact: high Difficulty: medium
O

First of all, you can make the account creation process much quicker by offering users the
D
SO

choice to sign up with Google, LinkedIn, Microsoft, or Facebook.

The only caveat here is that you have no control over submitted data (e.g., with a customs form,
W

you can exclude certain email providers, whereas 3rd parties won’t offer you that option).

As an alternative, you can abolish the sign-up process altogether by offering a guest checkout

for your e-commerce store. 

Here, you will face a potential trade-off between higher sales and richer marketing data. 

Decide for yourself if it makes sense to implement a guest checkout or alternative sign-up

options.

Source: Monday

48
42 Comparison landing pages
Category: layout Impact: high Difficulty: medium

I’m sure you know that competitors try to steal traffic from each other by bidding on competitor
brand name keywords in Google Ads or other PPC channels.

Yet, bidding on competitor keywords and writing unique ad copy is only a part of the process.

IN
Sadly, this is where most advertisers call it a day.

S.
You could achieve a much higher ROI from these ads by creating a custom landing page for
every big competitor.

AD
You don’t need to reinvent the wheel, but you have to show clearly that your company is

LO
superior to the competition. 

For instance, you can create a section with a table where you compare your features to other
N
W
products or services, thereby highlighting the advantages of your company.
O
D
SO
W

Source: Freshworks

49
Chapter VII: Things to avoid
Now that you’ve got a good basis and plenty of ideas to launch website experiments - there’s
something else that you have to know. 

There’s an extensive list of techniques that were heavily overused in the past and, in most
cases, will annoy your audience above all else.

IN
Note that there will be plenty of marketers that will disagree here, and that’s fine. The examples

S.
I’m about to present are usually frowned upon, but it doesn’t mean they won’t work for you.

In other words, you do you, but here’s my list of CRO techniques that you should avoid.

AD
LO
Browser notifications

Let’s be honest - in most cases, browser notifications are annoying and associated with spam.
N
W
Unless you have a very specific use case (such as news websites) - it's best to avoid them
O

altogether.
D
SO

Newsletter pop-ups

Perhaps, it's the most debatable point of them all. You’ve probably heard a lot of stories about
W

exit-intent popups increasing conversion rates, and that’s true - this can happen.

Yet, most popups (especially the ones that prompt you to subscribe to the newsletter) do more
harm than good.

Auto-play videos with sound

If you have a video anywhere on your page, make sure it’s mute by default. If you try to get
people’s attention on the web with sound - think twice.

People may be in the office, they might browse the web while their partner is asleep, and their
volume might be set to max - the last thing they want to hear is a video they never intended to
listen to.

50
Cheap traffic to build up a sample size

When you’re testing a new promising idea, it’s tempting to build up your sample size with cheap

traffic. Of course, push traffic or cold emails are much cheaper than LinkedIn or Google Ads, but

you’ll be sacrificing lead quality.

Run your tests on a targeted audience, and don’t chase high-traffic numbers just because it’s

cheap. You'll likely end up with thousands of random visitors and a few spammy form

IN
submissions.

S.
AD
Spin the wheel and other gambling-related techniques

Recently I’ve seen quite a few blogs with a spin-the-wheel game to win a free prize (like access

LO
to their private chat or a copy of their ebook). 

N
It's sad to see because most people associate spin-the-wheel and similar games with gambling
W
(and risk), and that’s probably not the brand image you wish to convey (unless you run an online
O

casino).
D
SO
W

My opinion: steer clear of any conversion tactics that are (even remotely) related

to gambling.

51
Chapter VIII: Analyzing the experiment’s

results

Now, before concluding your experiment and celebrating, you might want to do a little

reflection.

Let’s take a look at a few considerations from the statistics department that will help you judge

IN
the validity of your experiment’s results.

S.
AD
Did you attain an adequate sample size?

LO
The sample size is crucial, and here’s why. Say you have 20 visitors and 2 conversions on a

given day - that’s a conversion rate of 10%. The next day, your website gets 25 visitors and 6
N
conversions, which is roughly a 25% conversion rate.

W
Is it a 15% improvement? Of course not, because the sample size is too small. Those
O

conversions can be explained by a coincidence rather than a strong statistical pattern.

D
SO

I can’t tell you what the perfect sample size is because it varies based on niche and conversion

rate.

However, at least 2,000 visitors to each variation should be a good start.

Have your findings passed the “statistical significance”

test?

Statistical significance is a degree of confidence that your results were not achieved by

accident. It’s heavily reliant on sample size and the presence of outliers, among other criteria.

Now, most CRO tools will calculate this value for you, but there are also standalone calculators

designed for this purpose.

Usually, a 95% confidence level or higher is considered statistically significant, but it’s not set in

stone. You can set a lower benchmark (90%) or higher (99%) for your experiment.

I would advise you not to go any lower than a 90% confidence level, though, as in this case, the

chance of accepting a flawed outcome is too high.

52
Were there any outliers in your dataset?
You’ve got your dataset - now it’s time to get down to the nitty-gritty of analyzing it across

different variables.

For instance, you might find that traffic from a particular channel performed way better than the

rest, and adversely made your results positive.

Let’s say that most of your traffic comes from LinkedIn Ads and Organic Searches. You ran a

IN
test on both of these audiences simultaneously and discovered that visitors from LinkedIn Ads

S.
had a significant rise in conversions but not organic visitors.

AD
Applying your A/B test or landing page to the most promising traffic segment, as opposed to all

LO
traffic, makes more sense.Not only do you ensure that your effort is directed to the right place,

but you also remove the risk of outliers that skew your data.It can also work the other way
N
around - you see no noticeable increase in conversions and conclude that your initiative didn’t
W
work. What could’ve happened, however, is that a particular traffic segment had a huge drop in
O

conversions, leading you to conclude that the experiment wasn't successful.

D
SO

If you remove this outlier from your dataset - who knows, you might discover that your initiative

worked like a charm.

Thus, if you run a test across a broad audience to generate a high sample size, make sure to

break this data apart and look for individual patterns.

If needed, repeat the test for specific audience segments.

Was your test running for a sufficient time?

The ideal frame of an experiment is largely based on the number of days/ weeks it will take to

obtain statistically significant results. For high-traffic websites, it can be a week, while for low-

traffic sites, it can take months.

However, as a rule of thumb - run your experiment for at least 2 weeks, even if you can collect a

lot of data earlier. Giving your experiment more time to run means that you reduce some of the

random factors (day of the week, time, holidays) that otherwise would have affected the

outcome.

53
Conclusion

Here we are - you’ve reached the end of this book.

I hope that by now, you've learned all that it takes to launch successful and well-thought CRO

experiments and navigate your way through CRO jungles.

Whether you want to increase your sales, sign-ups, or engagement rate, I hope the practices

IN
presented in this book will provide enough guidance to succeed.

S.
Remember that conversion rate optimization is not a one-time project but a continuous process

AD
that requires constant testing, an open mindset, and dedication.

That's why I want to offer you my support in case you encounter any challenges along the way. 

LO
Feel free to reach out to me with any questions, and I'll do my best to help out.

N
Good luck on your marketing journey, fellow marketing explorer!
W
O
D

Thank you for reading this book, and if you’ve found it useful and enjoyable -
SO

please consider leaving a review. As you know, social proof is everything.

W

Footnotes

This ebook contains a number of links and mentions of 3rd party products. Be advised that

all of the recommendations or references related to 3rd parties have no commercial intent

behind them. The author has no affiliate or referral partnerships with any of the vendors

featured throughout the book and does not get compensated in any way or form

The screenshots of 3rd party websites presented in this ebook depict the layout of given

websites at the time of writing. As such, if you go to these websites now, the layout may

differ from the screenshots in this book.

54