The Defensible Forecast
How retail real estate teams get from gut instinct to a number they can stand behind in committee.
Publication Info
Your Art. Our Science.
Section 00
Executive Summary
Every new location is a multi-million-dollar bet, and someone on your team has to stand in front of a committee and defend it. Most site-evaluation tools hand that person a score and ask the room to trust it. Committees don't extend that kind of trust, and they shouldn't.
A single failed location can cost $450,000 to $1.2 million. The difference between teams that grow on plan and teams that stall isn't more data. It's whether the recommendation survives the question: where did this number come from?
This whitepaper examines why expansion recommendations fail in committee, and what a defensible forecast requires instead: scoring you can trace to its inputs, models calibrated to your own portfolio, honest ranges in place of false precision, and a feedback loop that makes every decision sharpen the next one.
Section 01
Where Recommendations Die
Every expansion decision ends in the same room. Someone presents a site. Someone else asks how they got the number.
Teams arrive at that number three ways today. A vendor's model produces a score nobody outside the vendor can explain. An analyst builds a spreadsheet nobody else can audit. Or experience calls it, dressed up in a deck. All three fail the same follow-up question, because the process behind the number was never designed. It accumulated: a mapping tool here, a demographics report there, a spreadsheet from 2019 that still does the real work.
How the Number Usually Gets Made
| Task | Typical Tools |
|---|---|
| Find available spaces | Listing platforms and broker emails |
| Check demographics | A GIS tool and census pulls |
| Analyze traffic | A foot-traffic subscription and DOT data |
| Build the financial model | Spreadsheets, in multiple versions |
| Track the pipeline | Another spreadsheet, or a CRM workaround |
| Share with stakeholders | Email attachments and drive links |
| Store signed documents | Shared drives |
| Communicate with brokers | Email, phone, text |
| Present to leadership | Decks built from scratch |
| Defend the number | Gut instinct and hope |
Not because these teams lack sophistication. Because no step in this workflow can explain the final number, and the person presenting it inherits that gap.
Three Ways the Number Fails
No trace
The score came from somewhere. Reconstructing where means reopening six tools and hoping the versions still match.
No reproducibility
Every analyst runs their own process. The same site can get two different numbers in two different meetings.
No downside case
A point estimate with no range invites the one question it can't answer: what happens if you're wrong?
The Stakes
A single failed retail location carries significant financial consequences. Even a buildout-only walk-away strands $450,000; ride out the full lease term and the exposure passes $1.2 million.
Failed Location Cost Model (Mid-Market, 3,000 sq ft)
| Cost Component | Amount |
|---|---|
| Buildout and tenant improvements | $450,000 |
| Remaining lease obligation (7 years) | $609,000 |
| Operating losses | $100,000 |
| Closure costs | $25,000 |
| Total exposure | $1.2 million |
Sources: Cushman & Wakefield 2025 Retail Fit Out Cost Guide, Statista Q3 2024, CBRE 2025 Retail Rent Dynamics
Scale Impact
For a 100-store retailer, even a 10% failure rate means 10 bad locations at roughly $800,000 each, the midpoint of the cost model above. The problem isn't a lack of data. It's disconnected data, with no way to defend the synthesis.
Section 02
The Anatomy of a Defensible Forecast
Defensibility isn't a feature you bolt on before the meeting. It's a property of how the number was made.
Operators have a word for the incumbent way: black box. A score arrives, confident and unexplained, and you're asked to bet seven figures on trust. A defensible forecast is built the opposite way, with four properties that hold up in the room.
Traceable, all the way down
Open any verdict and read the criteria your team wrote, the data sources it used, and the reasoning behind it. Open the forecast and see every variable and weight in the model. When a committee member asks why, the answer is a click, not a follow-up meeting.
Grounded in your portfolio
Projections modeled on your own stores and validated against markets you already know. A model that can't explain a store you've run for ten years hasn't earned a say in your next one.
Honest about uncertainty
A range with a midpoint and bounds, never a single brave number. Committees respect ranges. They punish false precision.
Aware of the rest of the fleet
A great site that pulls revenue from two existing stores isn't a great site. Cannibalization gets quantified against your actual locations before you present, not after you open.
The Committee Test
Before your next real estate committee meeting, ask yourself:
- 01Can you show the criteria and the data behind this site's score?
- 02Can you name the existing stores your forecast is modeled on?
- 03Do you know the downside case, not just the midpoint?
- 04Can you quantify how much revenue a new store would pull from your nearby locations?
- 05Could you answer the CFO's follow-up without leaving the room?
If any answer is no, the recommendation is running on trust. And trust is the first thing a committee runs out of.
Section 03
The GrowthFactor Approach
GrowthFactor was founded by MIT Sloan classmates who watched strong recommendations die for lack of a defensible number. The platform plans markets, evaluates sites, and manages deals in one place, and GrowthFactor Labs adds sales forecasts built on your own stores. Many teams run it as a second opinion alongside the tools they already have. It earns the seat by showing its work.
Market Planning & Site Intelligence
Market Planning
Know which markets fit before you shop for sites.
Match-based market mapping shows where conditions look most like the places you already win, so the pipeline starts in the right markets instead of the ones that happened to have a vacancy.
Custom Evaluators
Type an address. Read the verdict.
Your team writes each evaluator in plain language and picks its data: demographics, vehicle traffic, foot traffic, nearby businesses, cannibalization, and analog stores. It scores a site 1 to 5 with a written verdict you can open and check.
Sales Forecasts (Labs)
A range you can stand behind.
Labs builds a model on your own stores' sales history, then projects revenue for a new site as a range: midpoint, lower bound, upper bound. Once built, it forecasts a site in under 10 seconds.
Cannibalization
Know what a new store takes from the ones you have.
Trade-area overlap is quantified against your actual fleet, not industry hypotheticals.
Deal Pipeline & Collaboration
Deal Dashboard
Every deal, every stage, one view.
Kanban, table, and map views over the same pipeline, with stages from Searched through Grand Opening. Every deal stays on file with the numbers you had at signing.
Deal Pages
The whole deal history in one place.
Each deal keeps its files, custom fields, and an activity feed with comments and @mentions, so the next person can pick it up without a handoff meeting.
Studies & Export
A full site report in about ten seconds.
A full site report in about ten seconds. Studies export to PDF or Excel, with the Agent transcript attached as an appendix.
The Agentic Layer
GrowthFactor Agent
Ask your pipeline a question in plain language.
The Agent works over your own deals, models, and markets inside the platform, so the answer arrives before the broker hangs up.
MCP Integration
The forecast comes to where your team already works.
GrowthFactor launched the first MCP integration in site selection (April 2026). Score three sites from a Claude prompt in about 30 seconds.
Same Trace Everywhere
Every answer shows its work.
Every tool call the Agent makes is visible as a row, so each answer carries its trace: the inputs, the data sources, and the reasoning. Defensibility doesn't get lost in chat.
GrowthFactor Labs
A data scientist who built your model and knows your markets.
High-stakes decisions deserve dedicated expertise. Labs pairs your team with data scientists who build custom forecasting models alongside you, so you understand exactly how your model works and can explain it to anyone who asks. Not a black box you can't question.
Contact an AnalystSection 04
Every Decision Makes the Next One Better
The quiet enemy of most expansion programs isn't a bad tool. It's the blank slate.
In most workflows, every deal starts from zero. The analysis behind the last twenty sites lives in old decks and one analyst's head, and when that analyst leaves, it walks out the door with them. A defensible forecast compounds instead.
A store opens
Actual performance flows back into the model. The projection you defended gets graded against reality, and the model adjusts.
The analog pool deepens
Every opening adds a store your next forecast can be modeled on. The comparisons get closer to home with every decision.
The team compounds
A new hire can pull every prior analysis on day one and be productive in hours. Nothing about how you decide depends on who's in the room.
The Payoff
Customers surveyed in January 2026 report roughly 80% fewer underperforming locations, with forecast error about half the industry typical. That's what a forecast earns by learning your fleet one opening at a time.
Section 05
Results from Real Retailers
GrowthFactor customers get from a first look at an address to a full site report in about ten seconds: more sites through the funnel, and a defensible number on every one.
Results at a Glance
Case Study A
Cavender's
Trust You Can Explain
Challenge
The western wear retailer wanted to accelerate expansion, but not on the word of a model nobody could explain to leadership. The forecast had to be one their own team could interrogate and stand behind.
Results
“Other services hide behind black-box models that are hard to trust. The beauty of GrowthFactor is they make site selection incredibly simple, and give us clear unbiased recommendations.”
— Mike Cavender, Co-Owner and Head of Real Estate, Cavender's Western Wear
Case Study B
TNT Fireworks
Speed When It Matters
Challenge
With a selling season measured in weeks around July 4th, TNT has to evaluate hundreds of potential sites in a compressed window. Missing the deadline means missing the year.
Results
“Our business is very seasonal, and we have to manage evaluating hundreds of locations in a short time frame, while picking the best ones. GrowthFactor makes doing that simple.”
— Carson Anderson, Managing Director, TNT Fireworks
Case Study C
Books-A-Million
High Stakes, Short Clock
Challenge
BAM needed to evaluate roughly 700 sites from each of two bankruptcy auctions, fast enough to bid with confidence. GrowthFactor's team delivered full scoring and revenue forecasts inside 72 hours.
Results
“GrowthFactor accelerated our team’s new store site selection process, allowing the team to expedite the review of hundreds of potential sites as part of a complex bankruptcy auction.”
— Matthew Furnas, Vice President, Books-A-Million
Interactive
What is tool sprawl costing your team?
Adjust the sliders to see the hidden cost of a disconnected site selection workflow, and what you'd get back with one you can defend. Starting values are illustrative modeling assumptions; set them to your reality.
How we calculate this
Annual analysis hours
Hours recovered annually
702
That's 17.6 work weeks back for your team.
$105K
time savings
$2.4M
risk reduced
$2.5M
total impact
Based on industry data: $900K avg. failed location cost (Cushman & Wakefield), $150/hr fully-loaded analyst cost, 78% analysis time reduction, 60% failure rate reduction (GrowthFactor customer data).
Section 06
Getting Started
Every retailer's expansion journey is different. GrowthFactor meets you where you are.
See It on Your Own Markets
Bring an address you're actually weighing. We'll walk it end to end.
- Run your evaluator on a site you're weighing right now
- Open the verdict and trace every input
- See the pipeline and the committee-ready report
- Get pricing for your footprint
Start with Discovery
A 30-day analysis of your own sales history, free for brands with 40+ stores. No commitment beyond it.
- Share your sales and location data
- Learn who your customer really is and which variables drive your sales
- See your true trade area
- Keep the findings deck, whatever you decide next
Pricing
GrowthFactor Enterprise puts the platform in front of the whole team, and GrowthFactor Labs adds an embedded data science team. Current plans and details live on the pricing page.
See PricingAbout
GrowthFactor is the AI analyst for commercial real estate. Ask where to open next, whether a site is worth pursuing, or what a new store does to your existing ones. Watch it pull the data, run the analysis, and explain the answer in maps and tables. It does the analysis. You make the call.
Cofounded by MIT Sloan classmates Clyde Christian Anderson, Raj Shrimali, and Sam Hall. Backed by private and institutional investors including Teamworthy Ventures.
The projections and recommendations described in this whitepaper are based on historical data and analytical models. Actual results may vary based on market conditions and other factors outside GrowthFactor’s control. Survey figures are customer-reported (January 2026 customer survey).
Ready?
Walk into committee with the answer.
See a defensible forecast built on your markets and your own store data.