HOW TO BUILD A GOOD June 2015

COST MODEL Beth Zotter

 HOW TO BUILD A GOOD COST MODEL

AGENDA

 Bottom line numbers – CapEx, OpEx, $/unit

 Characteristics of a user -friendly model
 Data scavenger hunting tips
 Cost estimating tips
 Scenario modeling
 MODEL HEIRARCHY

Sue Babinec, ARPA-e “Tech to Market” presentation

PRODUCTION COST MODEL BOTTOM-LINE NUMBERS

Land /equipment
lifetime CapEx/yr /units/yr
Design &
Engineering
Depreciation,
($/yr)
Equipment CapEx/unit
OpEx/unit
Unit cost,
Labor
Installation ($/unit)
Fuel /units/yr + business costs :
Materials interest on loan,
taxes, etc.
(part of financial
CAPEX, OPEX, model)
($) ($/yr)
Also known as: capital expenditure, Also known as: operating
capital cost, investment cost, expenditure, operating cost
upfront cost
Hypothetical factory, simplified cost structure
CAPEX OVER CAPACIT Y: FOR COMMODITIES

Land
 Comparing options for producing
Design & same thing
Engineering  Rated output basis
 nameplate capacity
Equipment  Examples:
units/yr  $/kW
 $/MGY (million gallons per year)

Installation

CAPEX/capacity
CAPEX,
($/unit/yr)
($)
 IMPLICATIONS OF BOTTOM LINE NUMBERS

 Throughput Rate is key

 24/7 operation
Why factories
 “capacity factor” operate at night
= actual/potential output
80-95% for commercial,
less for pilot plant
 Know your rate limiting step!
 Include realistic downtime

 CapEx is more important than OpEx

 Time value of $:
 CapEx is now
 OpEx is in future and prices less certain
 So... Focus on lowering your CapEx fir st
YOUR MODEL CAN BE A TOOL FOR DECISION MAKING

This is a picture from Bad Piggies

 To be a tool where you design a go-cart to
navigate an obstacle course
 Unknowns are variables
 Physical system drives results
Design and
 “What if” redesign the cart
 High interdependency of to achieve target
parameters performance

 To drive decisions
 R&D effort is linked to cost drivers
 Sensitivities are known
 Actual, target, theoretical values defined
 Brutally honest
DEFEAT SKEPTICISM: SHOW YOUR ASSUMPTIONS

Show
assumptions!

REAL Modeled physical system COST

DATA
 WHY YOUR MODEL SHOULD BE USER -FRIENDLY

 For yourself

 For your team

 Everyone understands cost drivers

 Investors and partner due diligence

 Confusion  skepticism
 Visible logic  confidence

 Model complexity will grow, then shrink

 As real numbers replace calculations
 Streamlined as you weed out noise
 STRUCTURE THE MODEL SO ITS EASY TO FOLLOW

 Diagrams (make them)

 Process flow diagrams
 Device diagrams

 On separate tabs
 Physical system parameters
 Capital cost Summary
 Operating cost
CAPEX
 Summary page: key inputs/outputs,
financial assumptions OPEX
Physical (non-cost)
 Calculations Physical (non-cost)
 Clarify what numbers are inputs vs. calculated (color key)
 Easy to read : 220,000 vs 224531.692
 Show units everywhere
 Disaggregate calculations (show intermediate values) to avoid complex formulas,
which are hard to follow
 Name your variables for easy formula reading

 Comments (use them)

 CAPEX ESTIMATION

 Installed cost
 Generally ~3-5 times the cost of equipment and materials
 “Lang factor” in chem eng used as multiplier
 Can be much higher for pilot scale

 Scale factor s
 Use to estimate equipment cost by physical comparison
Cost B = Cost A*(Output B/Output A)^x
x is the scale factor, usually between 0.6 and 0.9
 Leverage known costs from other industries
 Quantify cost savings from process scale-up
Engine A Engine B
Caveats: 300 hp 150 hp
Min size – of industrial equipment available on market $50K $?
Max size – at a point, multiple machines are used instead
Commercial Your
model model
 Lifetime of equipment
 Impacts depreciated CapEx (total cost/lifetime)
 Maintenance schedule (downtime)
 STRUCTURE THE MODEL SO ITS EASY TO FOLLOW

 L e f t - to - r i g h t r e a d i n g
 O P E X , C A P E X s h e et s p u l l n o n - c o s t v a l u e s f r o m o t h e r s h e et s
 A s s u m p t i o n s e a s y to s e e a n d f o l l ow to s o u r c e

Physical calculations sheet (multiple tabs by module)

Module Parameter Unit Value Assumptions & source info

(all hypothetical)
OPEX sheet
Input Module units/yr $/Unit Total $ Assumptions & source info

Cost-driving physical
units drawn from
CAPEX sheet calcs sheets
Module Equipment Scaling unit Model Basis Basis Equipment Installed cost Installed
Value Value Cost Cost multiplier Cost
 INCLUDE A SIMPLE SUMMARY PAGE

 Clearly identify assumptions for scale and capacity factor

 Total CapEx, OpEx, and $/unit This model and others will be available in
 Show financial assumptions the group “Cost modeling resources”
folder on google drive
Example from Sandia Lab’s Reference Model Project for hydrokinetic devices
 TIPS FOR THE LONG SCAVENGER HUNT: WEB

 Market repor ts
 Free data only
 Price ≠ Cost

 Tip: screenshots!
 Google books
 Vendor specs
 Charts and tables
 because you won’t remember
your keywords

 “How its Made” videos

 Seeing process helps
parameterize your process &
equipment variables
 SCAVENGER HUNT: BROKERS AND VENDORS

 Brokers
 alibaba.com
 specialty equipment brokers for
industrial equipment
 ID the cost driving specification
 Examples: engine power, precision of
instrument, concentration, flow rate, etc.
 Quantify relationship between spec and
the cost
 Vendor s and OEM (original
equipment manufacturers)
 brochures
 salespeople
 Vendor quotes (custom equipment) –
 hard to get unless you mean business,
and have detailed specs
 very reliable if you can get them
 SCAVENGER HUNT: EXPERTS

 Probe with questions that elicit constraints (max/min):

- Location of bottleneck?
- Most variable input cost?
- Longest lead time item?
- Highest maintenance system?
- Age of oldest /newest equipment?

 Know process in advance

 Tour facility

 Pivot conversation if necessary

 SCAVENGER HUNT: EXCEL MODELS

 Getting information from other excel models, such as

 sources of data
 process assumptions
 installed cost multipliers
 input costs
 ...but always validate on your own. Models are by definition not
real, and often idealized

Example from BatPaC (battery performance and cost) model by

Argonne Lab
 DO REALIT Y CHECKS: VALIDATE

 Interdependency requires high use of formulas = accuracy risk

Validate calculated values with real numbers

 Reach outside your model for real data from other industries
 What others know, and will compare you against
 Powerful in validating your model
 Even if major differences exist, focus on analogies

 Real numbers - preferable to calculated numbers when you

have them

 Results too good to be true? then they probably aren’t

 WHEN TO USE SCENARIOS

 Consider scenarios when

 Qualitative system differences
 Pilot/commercial scale

 Pilot/Commercial scenarios
 Pilot - uses near term technical targets → market entry cost
 Commercial – long-term, external communication, the possible

 Example dif ferences

 Pilot uses outsourced inputs (materials, labor, leased equipment) that
become in-house at commercial scale
 Step-change cost reductions at scale
 Pilot is retrofit, commercial is new construction
POTATO CHIP FACTORY