Dr Neeraj Sanan

Digital Marketing Metrics Lecture 3

Media Metrics (contd.)
FREE PAID EARNED

Posts Boost posts Too much

Internet, social Virals
Website SEM, Display data
Digital
whatsapp/ call/ Internal EDM Database FOCUS
Virals
email

PR Advertising W-o-m, social ✓

TV / Radio

How to
get data
Newsprint/ OOH
Traditional PR Advertising W-o-m, social

Events &
Activation Speaker Sponsorships PR
Digital: Media vs Products

• Digital Products
• Our own products getting viewers & revenues (read profits)
• Digital Marketing
• Leveraging the Digital Media for promoting ourselves
• Direct Database Marketing (CRM: customer relationship management, EDM : Electronic
direct mailing)

• Digital Products need traditional media also

• Traditional Products need Digital media also
• Omnichannel Products need omnichannel media
Lading Page/ Website: Anchor for digital marketing

Google Analytics Audience Profile from different channels

- Set up Google Analytics Code 1. Demographics

- Set up Goals 2. Location
1. Will help define the journey of the person 3. Device
who filled the form
4. Average Time spent on Website
2. Demographic, location, Device, Source, 5. Bounce Rate
Medium of the person who filled the lead
6. Average Time spent on different pages
form
7. Visitors on the Website
This will help in
- Tracking Leads
- Tracking Traffic Source
Digital Marketing
Digital Marketing: Terminologies

• SEM : Search Engine Marketing; keyword marketing. The ‘paid’ URLs that show up on the
side of organic results, when a user is searching something
• SEO: Search Engine Optimizing/ organic search. The URLs that show naturally
• Display Ads: Banners, Position, Videos, Images
• Ratings: UV (unique viewers), PVs (Page Views), Time Spent
• Contextual Ads: Serving ads after crawling a users’ content consumption
• Remarketing: Cookie tracking people who are dropping out and reaching them again
Metric D1: Impressions, Pageviews & Hits

• Impressions: # of opportunities that have been presented to people to see an advertisement

• Pageviews: # of times a page has been displayed to a user
• Hits (#) = Pageviews (#) * Files on Page (#)
• Pages composed of many small files generate numerous hits per pageview
• Measuring hits, though previously popular, has largely been abandoned as hits are a
function of site design
Metric D1: Purpose is to assess web site activity

• Ideally pageviews counts pixels returned to a server, confirming the page

was properly displayed
• Pageviews should be recorded as late in page-delivery process as possible in
order to avoid inflated counts
• Major web sites have an ad server involved in presenting advertisements
• This technology to serve different ads to different visitors is widely
available
• A great marketing tool but given this we must confirm who saw what
Metric D1: Complications & cautions

• Measures of responses of a server to page & ad requests from users’

browsers, filtered to remove robotic activity & error codes prior to reporting
often include duplicate showings to same visitor
• Gross impressions may be a better term
• Measures do not take into account
• # of ads actually seen
• Quality of what is shown
• Whether message appeared to a specific, relevant, defined audience
• Whether those who saw page have any recall of the page/ad
Metric D2: Rich Media Display Time & Interaction Rate

• Rich media: Interactive media that allows consumers to be more actively engaged

• Average Rich Media Display Time (#)

𝑇𝑜𝑡𝑎𝑙 𝑅𝑖𝑐ℎ 𝑀𝑒𝑑𝑖𝑎 𝐷𝑖𝑠𝑝𝑙𝑎𝑦 𝑇𝑖𝑚𝑒 (#)

= 𝑇𝑜𝑡𝑎𝑙 𝑅𝑖𝑐ℎ 𝑀𝑒𝑑𝑖𝑎 𝐼𝑚𝑝𝑟𝑒𝑠𝑠𝑖𝑜𝑛𝑠 (#)

• Rich Media Interaction Rate (%)

𝑇𝑜𝑡𝑎𝑙 𝑅𝑖𝑐ℎ 𝑀𝑒𝑑𝑖𝑎 𝐼𝑚𝑝𝑟𝑒𝑠𝑠𝑖𝑜𝑛𝑠 𝑤𝑖𝑡ℎ 𝐼𝑛𝑡𝑒𝑟𝑎𝑐𝑡𝑖𝑜𝑛𝑠 (#)

= 𝑇𝑜𝑡𝑎𝑙 𝑅𝑖𝑐ℎ 𝑀𝑒𝑑𝑖𝑎 𝐼𝑚𝑝𝑟𝑒𝑠𝑠𝑖𝑜𝑛𝑠 (#)
Metric D2: Related Metrics

• Rich Media Interaction Time: Total amount of time a visitor spends interacting with
advertisement
• On same visit user might interact with the media twice
• Video Interactions: Video can be classified as rich media depending on way it is served to
viewer
• Typical interactions with a video: play, mute, pause
D2 Purpose: determine how an advertisement engages viewers

• Metrics track how long people spend viewing advertisement as a proxy for how interested
they are in its content
• A user who interacts with rich media is showing evidence of being more actively engaged
• Suggests they are more likely to take action, e.g., purchase
D2: Complications & cautions

• How to convert abundance of data into information?

• Marketers usually cut display times off at some point
• E.g., if rich media is displayed for more than X minutes, assume the consumer has been
distracted
• Averages might hide huge variances among users
• Be aware of changes in way the data was gathered!
• What counts as interaction? Depends on form of ad
• E.g. spending 1 second with mouse over the impression?
• Interaction is not an end in itself
Metric D3 : Clickthrough Rate

• Clickthrough rate measures impressions that lead to an action

𝐶𝑙𝑖𝑐𝑘𝑡ℎ𝑟𝑜𝑢𝑔ℎ𝑠(#)
• Clickthrough rate (%) = 𝐼𝑚𝑝𝑟𝑒𝑠𝑠𝑖𝑜𝑛𝑠(#)

• Note unless user clicks on “Buy Now”, clickthrough rate only measures one step on path to a
sale
D3: Complications & cautions

Large videos are likely to

increase # of people who
abandon process before ad is
served, especially if potential
customers have slow
connections
D3: Complications & cautions
• On sophisticated sites # of impressions ≠ # of pageviews as different ads are
served to different people
• Good news: details usually given in reporting package
• Assess absolute click numbers as well as % of impressions
• Clickthroughs count clicks not users
• Clickthrough rates for banner ads are often very low
• Identify ‘dead end’ pages visitors rarely move beyond
Metric D4: Cost per Impression, Cost per Click & Cost per Order

𝐴𝑑𝑣𝑒𝑟𝑡𝑖𝑠𝑖𝑛𝑔 𝐶𝑜𝑠𝑡 ($)

• Cost per Impression ($) =𝑁𝑢𝑚𝑏𝑒𝑟 𝑜𝑓 𝐼𝑚𝑝𝑟𝑒𝑠𝑠𝑖𝑜𝑛𝑠 (#)
Input
𝐴𝑑𝑣𝑒𝑟𝑡𝑖𝑠𝑖𝑛𝑔 𝐶𝑜𝑠𝑡 ($)
• CPM($) =𝑁𝑢𝑚𝑏𝑒𝑟 𝑜𝑓 𝐼𝑚𝑝𝑟𝑒𝑠𝑠𝑖𝑜𝑛𝑠 𝑖𝑛 𝑇ℎ𝑜𝑢𝑠𝑎𝑛𝑑𝑠 (#)

𝐴𝑑𝑣𝑒𝑟𝑡𝑖𝑠𝑖𝑛𝑔 𝐶𝑜𝑠𝑡 ($)

• Cost per Click ($) = 𝑁𝑢𝑚𝑏𝑒𝑟 𝑜𝑓 𝐶𝑙𝑖𝑐𝑘𝑠(#)
Output
𝐴𝑑𝑣𝑒𝑟𝑡𝑖𝑠𝑖𝑛𝑔 𝐶𝑜𝑠𝑡 ($)
• Cost per Order ($) = 𝑂𝑟𝑑𝑒𝑟𝑠(#)
Business vs Marketing: CAC vs LTV

• Ineffective or badly targeted ads may have a small CPM, but a high Cost per Click
• Companies prefer to compensate media & agencies on basis of new customers
acquired rather than impressions
• Agencies may push back as this arrangement passes risk onto them
• Acquisition depends on the quality of the product as well as the advertising
• Remember LTV: Life Time Value or
• CAC: Cost of Acquiring a Customer
𝐴𝑑𝑣𝑒𝑟𝑡𝑖𝑠𝑖𝑛𝑔 𝐶𝑜𝑠𝑡
= 𝑁𝑢𝑚𝑏𝑒𝑟 𝑜𝑓 𝑁𝑒𝑤 𝐶𝑢𝑠𝑡𝑜𝑚𝑒𝑟𝑠 𝑤ℎ𝑜 𝑀𝑎𝑘𝑒 𝑎𝑛 𝑂𝑟𝑑𝑒𝑟 (#)
Visitors < Visits < Pageviews < Hits
Metric D5: See, Click, Come

• Visits (or ‘sessions’ in Google Analytics): # of sessions on a site

• Can consist of a single or multiple pageviews
• To calculate visits one must establish a timeout period
• Subsequent requests from an individual who visited page count as new visits only after
timeout period elapsed
• Visitors (or ‘users’ in Google Analytics): # of unique people who visit
• Abandonment refers to unfinished shopping carts
• Shopping carts that don’t result in purchase
• Very common online
D5 (contd.): See, Click, Come

𝐶𝑎𝑟𝑡𝑠 𝑁𝑜𝑡 𝐶𝑜𝑚𝑝𝑙𝑒𝑡𝑒𝑑 (#)

• Abandonment Rate (%) = 𝐶𝑢𝑠𝑡𝑜𝑚𝑒𝑟 𝐵𝑎𝑠𝑘𝑒𝑡𝑠 𝐼𝑛𝑖𝑡𝑖𝑎𝑡𝑒𝑑 (#)

𝑁𝑢𝑚𝑏𝑒𝑟 𝑜𝑓 𝐶𝑜𝑛𝑠𝑢𝑚𝑒𝑟𝑠 𝑤ℎ𝑜 𝑃𝑢𝑟𝑐ℎ𝑎𝑠𝑒 (#)

• Conversion Rate (%) = 𝑁𝑢𝑚𝑏𝑒𝑟 𝑜𝑓 𝑉𝑖𝑠𝑖𝑡𝑜𝑟𝑠 (#)

𝐸𝑠𝑡𝑖𝑚𝑎𝑡𝑒𝑑 𝑆𝑡𝑜𝑟𝑒 𝑉𝑖𝑠𝑖𝑡𝑠 (#)

• Online to Offline Conversions (%) = 𝑂𝑛𝑙𝑖𝑛𝑒 𝐴𝑐𝑡𝑖𝑜𝑛𝑠 (#)
D5: Complications & cautions

• Visitors are more difficult to measure than visits

• Consumers often aren’t keen to register on different sites
• Users employing different browsers counted as different users
• To count unique visitors, companies often use cookies
• Files downloaded on one’s computer for future identification
• Cookies are also used to offer customized content
• Metrics can be distorted by automatic activity (“bots”)
• Bots aim to classify Web content
• Should eliminate bots IP addresses before reporting data
Metric D6 Bounce Rate (Web Site)

𝑉𝑖𝑠𝑖𝑡𝑠 𝑡ℎ𝑎𝑡 𝐴𝑐𝑐𝑒𝑠𝑠 𝑂𝑛𝑙𝑦 𝑎 𝑆𝑖𝑛𝑔𝑙𝑒 𝑃𝑎𝑔𝑒 (#)

• Bounce Rate (%) = 𝑇𝑜𝑡𝑎𝑙 𝑉𝑖𝑠𝑖𝑡𝑠 𝑡𝑜 𝑡ℎ𝑒 𝑆𝑖𝑡𝑒 (#)

• Measures site’s effectiveness in encouraging visitors to continue visit

• See more at Google Analytics Academy
• Money spent generating traffic for whom site is not relevant -- high bounce rates -- is money
wasted
• Use Content Experiments
• Show different pages to different visits and see which perform best, have the lowest
bounce rate
Bounce Rate: Complications & cautions

• Metric often reported by default by site’s host

• Visits (rather than visitors) likely to be used in computation
• Lowering timeout period will result in ↑ bounce rates
• Bounce rate can be defined for individual landing pages
• Some effective pages will have high bounce rates, e.g., directions
• Metric’s value depends on objectives of organization
Metric D7 Social Media Metrics: Friends/ Followers/
Supporters/ Likes
• Friends (#): # of friends of entity registered on a social network
• Indicates active interest in owner of the page
• Similar to Followers/Supporters etc…
• Likes (#): # of individuals favoring a social networking post/page
• Liking is a very low commitment activity
• Still more likes probably indicates greater appeal
Social Media Costs: Construction

𝑇𝑜𝑡𝑎𝑙 𝐶𝑜𝑠𝑡 𝑡𝑜 𝑃𝑟𝑜𝑣𝑖𝑑𝑒 𝑆𝑜𝑐𝑖𝑎𝑙 𝑁𝑒𝑡𝑤𝑜𝑟𝑘 𝑃𝑟𝑒𝑠𝑒𝑛𝑐𝑒 ($)

• Cost per Friend ($) = 𝑁𝑢𝑚𝑏𝑒𝑟 𝑜𝑓 𝐹𝑟𝑖𝑒𝑛𝑑𝑠 (#)

𝑇𝑜𝑡𝑎𝑙 𝐶𝑜𝑠𝑡 𝑡𝑜 𝑃𝑟𝑜𝑣𝑖𝑑𝑒 𝑆𝑜𝑐𝑖𝑎𝑙 𝑁𝑒𝑡𝑤𝑜𝑟𝑘 𝑃𝑟𝑒𝑠𝑒𝑛𝑐𝑒 ($)

• Cost per Like ($) = 𝑁𝑢𝑚𝑏𝑒𝑟 𝑜𝑓 𝐿𝑖𝑘𝑒𝑠 (#)

• Typical costs include design & update sites, devising social media marketing strategies

• Often very hard to attribute specific outcomes to social networking actions

• Yet a vital part of many marketing strategies
Social Media: Complications & cautions

• Some segments more reluctant to reveal brand loyalty

• 2 equally strong brands can have different levels of social presence
• Products used privately less likely to gain public support
• “Friends” & “Likes” intermediary metrics, not actual aim of organization
• Value of a like cannot be attributed solely to social media strategy
• Those incentivized to follow/ friend brand are likely to be less valuable than those who
followed without prompting
• Value is profit generated from customer, not revenue
Metric D8: Downloads

• Downloads (#): # of times that application or file is downloaded

• E.g., apps, software trials, ring tones, pictures, etc.
• Does not distinguish ‘1 individual downloading 10 times’ vs. ‘10 individuals downloading 1
time’
• Do you count completed or finished downloads?
• In standard analytics packages, downloads can’t be tracked like normal “pages”
Over to
mind-
measures