by Ambient Array
Find the next location your revenue history points to.
Revenue prediction for multi-location operators
Your strongest expansion opportunity may not be the busiest location or the largest market.
Tupll uses your existing location addresses and historical revenue to identify the neighborhood characteristics associated with your brand's performance, then predicts revenue at candidate sites. Compare a shortlist, or identify promising areas before the property search begins.
15+ Years
building site selection models
2,342
locations evaluated
124
verified builds/leases based on our reports
What Is Tupll?
Tupll is Ambient Array's site selection prediction service for multi-location operators. It uses your existing addresses and historical revenue to build a model of how surrounding consumer and business characteristics relate to your brand's performance. We apply that model to candidate locations, producing a revenue prediction for each site. Use the results to compare a shortlist or focus a broader market search.
Identify High-Potential Markets
Market planning that ranks and sequences the trade areas most likely to perform
Standardized Scoring
Compare candidate sites through a defensible scoring model
Reveal Hidden Growth
Growth pockets that traditional demographic tools miss
Prioritize Opportunities
Across an entire geographic footprint with confidence
How Site Selection Is Usually Done, and What Tupll Adds
Most site selection starts with a candidate property and asks whether it looks like a good opportunity. Teams may examine demographics, traffic, competition, comparable locations, broker recommendations, financial projections, and their own experience.
Tupll approaches the problem from the other direction. Instead of assuming which characteristics should make a location successful, Tupll analyzes actual location performance to identify which characteristics really matter. Those relationships can then be used to compare new opportunities and identify areas worth investigating before a specific property defines the search.
Traditional site selection
The Tupll approach
Starting point
Traditional
Start with an available site
Tupll
Start with where opportunity appears strongest
Evidence
Traditional
Review demographics, traffic, competition, and comparable locations
Tupll
Test which of those characteristics actually relate to performance
Method
Traditional
Apply thresholds, analogues, forecasts, and judgment
Tupll
Learn the relationships from actual location performance
The question
Traditional
“Does this location look good?”
Tupll
“What would we expect this location to produce?”
Next step
Traditional
Send promising properties through deeper diligence
Tupll
Focus real estate and financial diligence on the strongest opportunities
Why Tupll Works
A useful site model starts with your business, not a universal checklist. Tupll tests which neighborhood characteristics relate to revenue across your own locations, then uses that evidence to evaluate new ones.
Measured at Several Distances
We pull 40 to 50 consumer and business variables around each of your locations, at several radius bands, in both absolute counts and as a share of the surrounding area. How many of the right customers are nearby and how concentrated they are answer different questions.
Predictive Modeling, Not Static Data
Tupll learns from your own revenue history instead of a snapshot of today's demographics. A high income or a big population tells you nothing about what to do. A predictive score does.
Only the Variables That Earn Their Place
We pretest each variable individually against your own revenue, then pare the pool down to a small set that carries real, non-redundant signal. A variable does not survive because it sounds relevant. Two restaurants, or two retailers, never get the same model.
Several Models, One Prediction
We train several machine learning models on the variables that survived, then combine their outputs into one revenue prediction per site. You get a single number to compare, not a set of competing model results to reconcile.
Glass-Box, Not Black-Box
We show you exactly what the model includes and what it leaves out, and we share the underlying numbers. You see why a zone scored the way it did, not just a figure to trust on faith.
Tested Against Locations It Never Saw
We hold out a portion of your existing locations, predict them blind, and compare those predictions against what those locations actually earned. That comparison is how reliability gets judged, not how closely a model fits the history it learned from.
Our Site Selection Methodology
We start with your location history, test the surrounding market data, and build a model that predicts revenue at candidate sites. Here is how the analysis works.
Start with your locations and revenue
You provide existing location addresses and historical revenue by location. Those are required inputs for the machine learning model. We then query 40 to 50 consumer and business variables around each location, using sources including Acxiom. Measures are collected across multiple radius bands and expressed as both counts and proportions. The same characteristic can mean different things close to a location and farther away, so we test which measures at which distances are associated with your brand's revenue. A count tells you how many households or businesses are nearby. A proportion tells you what share of the surrounding market they represent. Tupll tests both against your historical revenue, so a large market does not automatically outrank a smaller market with a more relevant customer mix.
Keep the signal. Remove the overlap.
First, we test each candidate variable individually for correlation with your historical revenue. We then reduce the pool to a small set that carries useful, non-redundant information. A variable must contribute more than an interesting description of the neighborhood. It must help the model distinguish locations by revenue without merely repeating information another variable already provides.
Combine several models into one revenue prediction
We train several machine learning models on the reduced set of variables. Their outputs are combined into one final revenue prediction for each candidate site. The combined model is then evaluated against actual revenue at locations withheld from training.
Compare sites by predicted revenue
Each evaluated site receives a revenue prediction based on your brand's historical performance and the surrounding consumer and business characteristics. We use those predictions to rank the evaluated sites. The prediction is not a combined score for property feasibility, profitability, and strategic value. Those remain separate parts of your investment decision.
Test predictions against locations withheld from training
We hold out a portion of your existing locations, predict their revenue, and compare those predictions with actual results. The errors help us evaluate and refine the model. A check used to guide model revisions is part of development. It should not be confused with an untouched final test, and a strong historical result is not a guarantee of future revenue.
Executive Interpretation Layer
We translate the model's outputs into clear recommendations aligned with your expansion strategy, budget, and internal decision patterns.
See How Tupll Works
A short walkthrough of our site selection methodology, from raw data signals to decision-ready location scores.
Who Does What?
Tupll evaluates the surrounding market. Your real estate team evaluates the property. Your finance team evaluates the investment. Here is how the responsibilities fit together.
Tupll's Role
Our experience has shown us that the neighborhood around a site explains most of how well it ultimately performs. Our site selection services cover these functions:
- •Trade area delineation
- •Neighborhood mapping
- •Neighborhood statistics gathering
- •Demand-side indexing
- •Site suitability scoring
- •Catchment analysis
Your Real Estate Team's Role
Meanwhile, we leave the "last mile" analysis to your team or other external partners, including:
- •Parcel-level feasibility
- •Zoning and entitlement review
- •Environmental and Geotech flags
- •Infrastructure readiness assessment
- •Utility load calculations
- •Civil engineering pre-con review
Your Finance Team's Role
Finally, your financial team will be in charge of these important tasks:
- •Sitework cost estimation
- •Pro forma cost modeling
- •CapEx forecasting
- •Budget-to-build alignment
- •GC bids
- •GC value engineering
- •Critical path scheduling
Tupll measures the demand around a site. Your team decides whether a particular property can capture it. Your team's role is to make sure specific properties, buildings and facilities fit your needs, both technically and financially. By taking the neighborhood analysis piece off of your hands, you can free your time and resources to carry out these other very important activities.
We never tilt you toward a specific building, and we do not even check what is available, so there is no conflict of interest. We see things in the data that are hard for you to see. You know your brand. Both matter, and we keep those roles separate on purpose.
What a Tupll Project Looks Like
Your team provides the location history. We build the model, evaluate candidates, and walk you through the findings.
Data Intake & Environment Modeling
We analyze the revenue performance of your existing locations against key success-correlated consumer and commercial factors around each one.
Model Development & Calibration
We build you a 100% custom Tupll site selection model, incorporating a blend of multiple top machine learning models using supervised learning.
Market & Site Scoring
We put your new model to use, evaluating candidate locations along the same dimensions as your model.
Executive-Level Insights Delivery
You receive a revenue prediction score, key location statistics and heat maps for each evaluated location.
Advisory & Ongoing Support
We partner with your team to interpret findings so that your team can secure available land and facilities opportunities within your top revenue location zones.
A revenue prediction, with the analysis behind it
Revenue prediction for each evaluated site
Compare candidate locations using a model trained on your own location history.
Ranked comparison
See how the evaluated sites compare by predicted revenue, whether you are assessing a shortlist or exploring a market.
Supporting location data and maps
Inspect the consumer and business measures around each site, including the radius bands used in the analysis.
Model explanation and recommendations
Understand which variables the model retained, why sites ranked differently, and what the findings mean for your search.
Tupll evaluates neighborhood-level revenue potential. Property availability, access, feasibility, costs, and operating plans still need your team's review.
How Pricing Works
The model build and candidate-site evaluations are priced separately.
One-time model build:
Custom revenue model
For brands with existing locations and historical revenue by location. Includes the statistical analysis and machine learning model development.
First-location analysis
For brands without location revenue history. Uses statistical and GIS analysis without supervised machine learning.
Candidate-site evaluations:
Minimum batch of nine locations. Site-evaluation fees are additional to the model-build fee, so the minimum evaluation batch is $1,080.
Starting investment for a brand with existing locations:
A $15,000 model build plus nine site evaluations. Evaluations are purchased in batches of at least nine, even if one candidate site is your immediate priority. For a market-planning project, contact us to discuss the number of evaluation locations needed.
* Prices are subject to change without written notice. Confirm current pricing and scope with our customer success team before commissioning work.
ROI Calculator
Explore how an assumed improvement in location revenue would affect five-year profit and ROI. This calculator illustrates a scenario, not a forecast or a measured Tupll result.
One-time model development fee.
$120 per location (paid per batch). Minimum batch size 9.
You can paste values with $ and commas.
Example: 20 = 20% profit margin.
The prefilled 15% is an illustrative assumption, not a published Tupll performance result. Replace it with the assumption you want to test. Revenue improvement is converted to profit using the margin you enter.
Results
5-year horizon- Model fee
- $15,000
- Evaluation fee
- $1,080
- Total investment
- $16,080
- Incremental annual profit
- $—
- 5-year incremental profit (return)
- $—
- Net gain (return − investment)
- $—
- ROI, First 5 Years
- —
- Estimated payback period
- —
Keep your tools. Find more places worth pursuing.
Your existing tools and team can help assess an opportunity. Tupll can also help find it. Use your location history to identify promising areas before availability defines the search, then bring your brokers, analysts, and finance team into the locations that merit a closer look.
What most teams already use, and keep using
Foot traffic & location intelligence
Placer.ai, Unacast, Cuebiq
Demographics & segmentation (GIS)
Esri Business Analyst with Tapestry, Claritas PRIZM, Experian Mosaic, Spatial.ai
Predictive site-selection platforms
SiteZeus, Buxton, Kalibrate, SiteSeer, eSite Analytics, Tango
Industrial & plant siting
Site-selection consultants, brokers (CBRE, JLL), state and local incentives, laborshed analytics (Lightcast, JobsEQ), and utility feasibility studies
The classics
Broker comps, demographic reports, drive-time and radius rings, pro-forma spreadsheets, and hard-won gut
A model you can use beyond the first decision
A site-selection consultant may advise on the market, the property, and the transaction. Tupll's engagement centers on a model trained on your location history, which can be applied to additional candidate sites as you expand.
Use it to direct a new market search or compare an existing shortlist. Your real estate and finance teams retain responsibility for property diligence and investment approval.
Build the model once. Evaluate new opportunities as you grow.
The $15,000 model-build fee creates a model trained on your brand's history. Candidate evaluations cost $120 per location, purchased in batches of at least nine.
That structure lets you apply the model beyond the first shortlist, including to areas your team might not otherwise investigate.
Give your search more possibilities, then compare them on predicted revenue.
Building Your First Location?
Our flagship modeling methodology involves analyzing your existing locations. However, we also build site selection models for brands building or leasing their first property. Ask us how we do it!
Who We Serve
Tupll is built for brands that open on a plan, not on whichever space comes available first.
Multi-Location Consumer Brands
Retailers, healthcare groups, wellness & fitness brands, and service networks expanding into new trade areas.
Commercial Real Estate & Development Teams
Organizations that require predictive clarity when evaluating complex markets and high-value projects.
Growth-Stage Organizations
Companies with evolving product mixes and emerging geographic footprints that need structured market intelligence.
Our Ideal Clients
Ambient Array's focus is on evaluating the neighborhood around your candidate locations to assess revenue potential based on area factors. We work best with clients who employ internal teams who are prepared to connect with local market realtors or to inquire directly about available properties.
Tupll is the right fit for multi-location operators who already have a few years of revenue history and treat each location as a serious investment. It is not the right fit when you have already chosen the location and the decision is closed, since there is nothing left for a prediction to inform. And it does not replace the tools you already use. Layer it on top of them as a more rigorous layer of intelligence.
Client Results & Testimonials
Over 15+ years we have evaluated 2,342 locations, and 124 builds or leases have been verified as going ahead on the strength of our reports. The first number is the scope of the experience. The second is evidence that clients acted on it. Neither is a measure of how closely a prediction matched what a location later earned, and we would rather say so than let the two be confused.
Jed Jones is one of the most talented machine learning professionals I've had the opportunity to work with. The solutions he delivered fundamentally changed how we sourced and evaluated new properties for expansion. His ability to uncover actionable insights from complex data created measurable business value and significantly improved our decision-making process.
Mark L.
Director of Data & AI, National Building and Roofing Solutions Company
We have engaged Jed and his team several times over the past seven years. The work he has done for us has included greenfield/new market analyses, as well as analytics that helped us determine whether we should purchase a competitor's existing location. We highly recommend Jed and his team for site selection and geographical market analysis projects. The investment has paid for itself many times over.
Victor S.
Owner & CEO, Roofing and Home Solutions Company Operating in Midwest/Northeastern US
I enlisted Tupll services when I was evaluating whether to purchase a new chiropractic location from the franchiser. I was able to evaluate multiple zones within my chosen market and I ended up choosing one in the Tupll-recommended zone. This turned out to be a very wise investment. The location was built, grown and then resold later to another franchisee for a very hefty profit.
Dustin S.
Investor and Multiple Business Owner
My client manages a multi-tenant building located in a strip mall that had a hard-to-fill vacancy. With their analytics approach, Tupll was able to recommend the appropriate type of business to go into that vacant unit. On that advice, we executed our prospective tenant outreach activities around this target industry type. Having that report in hand made the lease happen much, much faster and helped us anchor the right client. The tenant has been thriving in this location ever since.
Francoise L.
Agency Owner
Straight Answers to the Questions We Hear
Before signing on, most teams push back with the same few questions. Here are honest answers.
What does Tupll add if we already know our markets?
Your team knows the brand, the properties, and the local conditions. Tupll tests another source of knowledge: the relationship between your existing locations' revenue and the markets around them. The value is not simply confirming what you already believe. It is finding a promising area you have not considered, or showing where a familiar rule does not hold across your own portfolio. Your team brings that finding back to the property and the deal.
Isn't this just napkin math, a radius with a logo on it?
No. A radius report describes the surrounding area. Tupll tests 40 to 50 consumer and business variables across multiple radius bands, using both counts and shares. Each candidate variable is screened against your historical revenue, the pool is reduced to useful, non-redundant variables, and several models are trained and combined into one revenue prediction per site.
How accurate is it?
Accuracy must be evaluated for the model built on your data. The relevant evidence is how predicted revenue compares with actual revenue at locations withheld from training, together with the way those locations were used during development. Tupll predicts revenue from historical performance and surrounding market characteristics. It does not guarantee future revenue or account for every difference in property suitability, operations, or market conditions.
About Tupll
Tupll is a product of Ambient Array. We build site selection models for operators deciding where to open, build, lease, or acquire their next locations. Our work combines statistical analysis, machine learning, and interpretation of the results. The aim is practical: help your team identify promising locations and understand the evidence behind the comparison.
Get Started with Tupll
Ready to make confident expansion decisions? Contact us today and catch us up on where you are in your site selection journey. Fill out the form below or call: 512-521-6791 during business hours (9AM - 5PM CT).
Find more places your brand could perform well
Tell us where you want to expand and what location revenue history you have. We can discuss whether Tupll's modeling approach fits the decision, from exploring a market to comparing candidate sites for an opening, lease, build, or acquisition.