How We Actually Make Location Decisions: A Six-Phase Framework for Corporate Site Selection

There is a widely held belief in corporate real estate circles that site selection is mostly relationships and intuition — that the right consultant calls the right contact at the right utility, and the deal materializes. That version of site selection exists, and it produces mediocre outcomes about as often as it produces good ones.

The version we practice is a structured data process. The intuition matters — fifteen years of walking sites and negotiating incentives teaches you things you cannot get from a spreadsheet — but the intuition is a finishing layer on top of a rigorous analytical foundation, not a substitute for it.

That foundation maps almost perfectly onto the six phases of data analysis: Ask, Prepare, Process, Analyze, Share, and Act. What follows is how each of those phases actually plays out when a company is deciding where to put a half-billion dollars.

Before You Touch the Data, You Have to Ask the Right Question

Most engagements that go sideways do so in the first two weeks. Not because the data was wrong or the analysis was flawed, but because nobody took the time to define what success actually looked like.

“We need a new manufacturing facility” is not a question. It is a project. The question underneath it might be: “Which labor markets in the Southeast can support a 400-person advanced manufacturing operation with a $50M annual payroll, given our 2027 start-of-production requirement?” Or it might be: “Where can we site a 200MW data center campus with long-term power certainty within 18 months?” These are different problems. They pull different datasets, weight different variables, and produce different answers.

“Every ambiguity you leave in place here becomes an argument at the finalist stage, when you have already spent six months and several hundred thousand dollars narrowing the field.”

The Ask phase forces a company to externalize assumptions they have often never written down. What is the minimum viable workforce size? What is the actual power requirement at full buildout, not just Phase 1? Is the $80M incentive target a hard floor or an aspiration? Does the CFO's “Midwest preference” mean the entire MISO footprint, or does it mean Indiana and Ohio?

Prepare: You Cannot Analyze Data You Do Not Have (Or Cannot Trust)

Once the question is clear, the next challenge is assembling the inputs. This sounds administrative. It is not.

A typical corporate site search requires data from no fewer than a dozen sources: utility interconnection queue data, labor market statistics, road and rail infrastructure assessments, environmental databases, real property records, incentive legislation, and tax rate schedules, among others. Each of these has its own cadence, its own formatting conventions, and its own failure modes.

Utility data is a good example of why preparation matters. Published transmission capacity figures are often months or years out of date. Grid interconnection queues — which tell you whether that 200MW is actually deliverable in your timeframe — are dynamic and require direct utility engagement to interpret accurately. A site that looks power-certain on a static map may have two years of queue position standing between it and your switchgear.

Good preparation means knowing which sources to pull, how frequently they are updated, what their known gaps are, and — critically — which figures require primary source verification versus which you can accept as reported. We maintain a standing protocol for this. Not because it makes us look organized, but because the alternative is discovering a data gap in week ten of a twelve-week search.

Process: The Unglamorous Work That Determines Whether Your Analysis Is True

Data cleaning is where most amateur site selection analysis breaks down, because it is the phase that requires the most discipline and produces the least visible output.

Consider workforce data. Lightcast and Chmura both report occupational employment statistics, but they use different methodologies, different geographies, and different vintage years. Pull them both into an analysis without harmonizing the definitions, and you will find yourself comparing different things while believing you are comparing the same thing. The labor market that looks competitive for a semiconductor fab on one platform may look marginal on the other — and both may be describing reality from a different angle.

Processing means bringing data to a common standard before you run a single comparison: consistent geographies, consistent time horizons, consistent unit definitions, and a documented record of every transformation applied.

This phase is also where we conduct the environmental baseline — verifying that the sites under consideration do not carry latent liability. NFA letters, PFAS baseline conditions, hydrology checks. The due diligence investment here is small relative to the cost of discovering a contamination issue after you have already announced the project.

Analyze: Where the Data Tells You Something, If You Let It

With clean, harmonized data in hand, the analysis phase is where location decisions actually take shape. The mechanics — weighting models, scoring frameworks, sensitivity analysis — are less important than the question you are trying to answer with them.

A weighted scoring model is a communication tool as much as an analytical one. It makes the decision criteria explicit, forces tradeoffs into the open, and creates a defensible record of how you arrived at a recommendation. What it does not do is replace judgment. When three finalists score within two points of each other on a 100-point scale, the model has told you that they are operationally equivalent.

The decision then turns on factors the model cannot capture — the quality of the regional workforce commission, the track record of the state economic development agency on incentive delivery, the infrastructure commitment the utility is willing to make in writing versus what they said on a call.

Visualization matters here too — not for aesthetics, but for communication. A map of finalist communities with power availability indexed against labor cost per job is a fundamentally different artifact than a table of the same numbers. One is for analysis. One is for presenting findings to a board that has twelve other items on its agenda.

Share: The Recommendation Is Only as Good as the Explanation

The Share phase in site selection is the deliverable — the formal recommendation and the reasoning behind it. It is also, in our experience, the phase most consultants underinvest in.

Corporate location decisions involve a remarkable number of stakeholders whose buy-in matters: legal, environmental, finance, operations, HR, government relations. Each has a different set of concerns. Each will ask different questions. A recommendation that the operations team finds compelling but that the CFO cannot defend to the board is not a final recommendation — it is the beginning of another round of analysis.

Our deliverables are built with this in mind: an executive brief that speaks to the strategic rationale, supporting documentation that can withstand technical scrutiny, and an incentive summary accurate enough to inform negotiation without creating liability if terms change. The analysis does not end when you identify the winner. It ends when everyone who needs to act on the recommendation understands why it is the right call.

Act: Putting the Recommendation to Work

The Act phase is where site selection typically gets handed off — the consultant delivers the recommendation, the company executes the decision, and the project moves into construction. That is one version of the relationship.

The more valuable version keeps the advisory function engaged through the critical windows where decisions made in haste become costs absorbed for a decade. Incentive negotiations. Infrastructure commitment timelines. RFP language. The specific milestones tied to job creation thresholds that determine whether the tax credits are 80% recoverable or 40% recoverable.

We have seen projects where the location decision was correct but the execution was costly because the negotiating team did not know what the economic development organization was actually willing to concede. The site was right. The deal was not optimized. Those are different problems, but they are connected.