How to Analyze a Business with Incomplete Data (Real Case)

How do you value (or even understand) a business when the numbers are incomplete, inconsistent, or simply unavailable?

A few years ago, I was asked to value a mid-sized regional distributor for a potential acquisition. No audited financials. No CRM data. No formal board reporting. No cash flow statement.

Although the target’s management was cooperative, the information flow was chaotic. They provided three years of internal P&Ls, but the cost allocation policy had changed recently; additionally, the balance sheet had several intercompany loans with no documentation or reconciliation. A recipe for disaster.

Most analysts would stall. We didn't.

The gap between stalling and proceeding with what you have is exactly what this post is about.

After several years in the field, we understood the harsh reality: you will never have all the data. Not next quarter. Not after the due diligence room closes. Not even if you are the CFO.

Waiting for perfect information or proceeding with what we got?

In practice, most decisions happen upstream of an ideal scenario with clean and detailed financial statements: early-stage companies, cross-border opportunities, informal markets, and, of course, insufficient disclosures.

In those contexts, the analyst’s job is not to wait for perfect information; it is to impose structure on imperfect, or insufficient, inputs, which requires a shift in mindset: from “measuring precisely” to “reasoning probabilistically.”

The issue is that most analysts don't have a structured approach to operating in the absence of data. They either over-rely on whatever data exists (and miss the distortions), or they freeze waiting for information that will never arrive in the form they expect.

What we actually need is an analytical posture: a disciplined way of structuring ambiguity, extracting signals from informational clutter, and arriving at defensible judgments even when the picture is incomplete.

The real question then becomes: What can we still know, and what needs to be assumed?

Case: Evaluating a mid-sized distribution business with partial visibility

Back to the regional distributor. We were asked to assess a business operating in a fragmented sector. The available information looked something like this:

• Revenue figures: partial and inconsistent across years

• No formal financial statements (only internal P&Ls)

• Tax returns (that did not match any of the information above)

• Anecdotal evidence of growth

• Operational data scattered across conversations, not systems

At first glance, very difficult to use. But analysis does not start with data. It starts with structure.

Here is the framework we applied:

Step 1: Identify the structural question first

Before touching a single number, we asked: what would need to be true for this business to be worth backing? That reframing is critical. It shifts the analyst from data-gathering mode into hypothesis-testing mode. Instead of cataloguing what we didn't know, we were building a map of what we needed to verify.

In this case, the business had to demonstrate consistent cash generation above debt service, and client concentration below a threshold that would make key-account loss unacceptable to the buyer. Those became our two organizing questions.

Step 2: Reconstruct the economic engine

Instead of asking “How much is this business worth?”, we asked:

• What drives revenue? (volume × price)

• What drives cost? (fixed vs. variable structure)

• Where are the bottlenecks?

We mapped the business like a machine. Inputs → processes → outputs.

Even rough estimates begin to anchor reality.

Step 3: Identify invariant relationships

Certain relationships hold even when the data is messy.

Businesses are economic systems before they are accounting systems. Financial reporting may be incomplete, delayed, manipulated, or inconsistent, but the underlying mechanics still leave traces. Growth consumes working capital. Low-margin businesses require scale. Operational complexity eventually shows up in headcount, logistics, or customer concentration. Reality leaks through structure.

This fact becomes extremely useful when the reported numbers themselves cannot be fully trusted.

• Margins cannot expand indefinitely without a justified operational change.

• Sustained growth must come from somewhere: more customers, higher ticket sizes, geographic expansion, improved pricing, or product mix shifts (in this case, pricing power was severely limited by the industry structure).

• Cash constraints eventually surface in receivables, supplier pressure, or working capital deterioration.

Management claimed that revenue had nearly doubled over a relatively short period. Fine. Then what absorbed the operational load?

• Did warehouse capacity expand?

• Did logistics costs rise?

• Did headcount follow?

• Did supplier terms tighten?

• Did pricing materially improve?

The operating footprint and the reported growth trajectory were not fully reconciling. The business was telling one story operationally and another financially. That gap became a central area of investigation.

Analysts often treat financial statements as isolated outputs rather than expressions of an underlying economic machine. But even imperfect businesses leave operational fingerprints. The task is learning how to read them.

Step 4: Use triangulation instead of precision

When direct data is unreliable, indirect signals matter more:

• Market benchmarks (industry margins, growth rates)

• Client behavior (is A/R healthy?)

• Client concentration (dependency risk)

We built ranges, not points. Not “EBITDA is 18%,” but “EBITDA is likely between 12% and 20%, given observed constraints.”

This reframes the output from certainty to bounded judgment.

The tax returns showed revenues 11% below what the internal P&Ls indicated. That alone is a red flag.

We dug into the bank statements. Cash deposits corroborated a revenue figure closer to the internal accounts. The owner’s explanation: he ran conservative figures for tax purposes but “knew the real number”.

That's not necessarily an intentional misrepresentation; it's the informal accounting reality of thousands of private businesses. However, it requires a specific analytical response: triangulate across sources, weight each by reliability, and build a range rather than a point estimate.

Step 5: Stress-test the narrative, not just the numbers

Most weak analyses fail here.

They accept the story and refine the spreadsheet.

We do the opposite.

We challenge the story:

• What must be true for this business to sustain its current trajectory?

• Where is the fragility?

• What breaks first under pressure?

Incomplete data increases the importance of this step. Numbers only tell part of the story. The conversation with the owner was the most informative data source we had.

How he talked about his clients (names, relationship history, payment behavior). How he described his team. Whether his explanations of past difficulties were self-aware or defensive. These are imperfect inputs, yes, but they were systematically gathered and weighted.

The concept here borrows from what good intelligence analysts call “source reliability weighting,” which is simply calibrating soft data (as opposed to ignoring it).

Case: Evaluating an early-stage software startup

We have applied this framework across different firms and industries; the result is always the same: a structured view of the risks, boundaries, and narratives underpinning any business, despite the absence of complete information.

Another example: we were engaged to value a software startup with only 14 months of operating history. No three-year trend and, of course, no audited numbers.

Instead of walking away, we identified the two truly unknowable variables: customer churn rate and sales efficiency (note that in this case, we put our attention not on what we got but on what we missed). We ran scenarios.

At 5% monthly churn, the valuation was USD 4 million. At 2% churn, USD 11 million.

That wide range was not interpreted as a model failure or as a freezing threat; it was used as a signal to structure the deal with performance-based earn-outs tied to churn. The client did exactly that.

Two years later, churn came in at 3.5%. The final price landed near the midpoint. Everyone understood the logic because we had mapped the uncertainty from the start.

Better data, better decisions?

There is a quiet misconception in finance (and business in general): that better decisions come from better data.

Sometimes they do.

But more often, better decisions come from better thinking applied to ambiguous situations with imperfect data.

In our experience, analysts trained only in clean environments struggle when ambiguity enters the room. They freeze. Or worse, they overfit fragile inputs into precise-looking outputs.

That is where judgment matters. And judgment is built by repeatedly structuring chaos.

We’ve seen two “failure modes” in financial analysis, and they’re rooted in the same cause:

The analyst who trusts the model

Give them clean data and they will produce a convincing output by missing every distortion hiding beneath the surface. The model becomes a substitute for judgment.

The analyst who never commits

This analyst distrusts incomplete data so deeply that they never commit to a view. They hedge everything, qualify everything, and ultimately contribute nothing to the decision at hand.

The standard we should hold ourselves to is different: structured confidence under uncertainty. It means knowing which variables drive the outcome, stress-testing the assumptions those variables rest on, and being explicit about where the analysis is fragile.

Navigating incompleteness

The ability to operate with incomplete information is not a workaround; it is the core skill of the industry.

Markets are not complete information environments. Neither are private businesses, acquisition targets, credit applicants, or expansion opportunities. The analyst who waits for perfect data will always be waiting.

What separates sound judgment from paralysis (or worse, from confident error) is structure. Structure the question first. Triangulate across available sources. Build scenarios rather than point estimates. Calibrate qualitative signal alongside quantitative data. And finally, be explicit about where you're uncertain and why.

A decision made with explicit, bounded uncertainty is often more reliable than a decision made with false precision.

You don’t need complete information to produce valuable insight. You need:

• A clear structure of how the business works

• A disciplined way to test relationships

• The ability to think in ranges, not points

• And the willingness to challenge narratives

The reality, however, is that incomplete information does not reduce the role of analysis; it shows what analysts are made for.

Have you faced an analysis where the data was incomplete or unreliable? What was the variable that ultimately drove your judgment call?

Drop your answer in the comments, or if you're working through a situation like this, let's talk about building the right framework for it.

Missing data is the default condition in private markets. The problem is understanding what can still be known… and what cannot.

Contact

contact@emporioadvisory.com
(+57) 310 530 0586

Social Media

Join our community of entrepreneurs and receive exclusive content

© 2026 Emporio Financiero. All rights reserved.

Spanish version