Bottom-Up Market Sizing for Untracked Niche Segments

Bottom-up market sizing builds the number from the ground up: a verifiable count of potential customers, multiplied by an evidenced estimate of what each spends per year, then cross-checked against a narrowed top-down figure. The arithmetic is deliberately simple. One standard expression is the total number of customers multiplied by annual contract value, segmented by customer type rather than blended, and bottom-up sizing carries credibility with investors precisely because every input is explicit and independently challengeable.

Top-down desk research breaks down in niche markets for a structural reason: analyst definitions are built to aggregate, not to isolate. A category report bundles the submarkets you do not serve with the one you do, and its sizing often rests on survey data rather than observed purchase behavior. The definitional slack is enormous. In a July 2025 release, MarketsandMarkets put the global cybersecurity market at USD 227.59 billion in 2025, while Precedence Research values the same market in the same year at USD 301.91 billion, because their category boundaries differ. A loose 'cybersecurity market' label can support almost any number across a wide band; a niche segment definition cannot be allowed the same freedom.

• The market definition is yours to write, and a vague boundary is where sizing fails before any arithmetic.
• There is no report to buy, so the evidence burden shifts from the analyst to the deal team.
• Every input must be sourced, dated and recomputable, because the investment committee will recompute it.

The stakes are transactional. A market claim that survives diligence protects the price; one that fails becomes a negotiation lever used against you. That asymmetry is why disciplined teams treat the sizing exercise as an evidence problem rather than a modelling problem, and why the rest of this article is organised around the inputs rather than the formula.

The count of potential buyers can be assembled from free and moderate-cost sources before anyone buys a report. The goal is a universe you can defend line by line, not a big number.

• Census business counts, filtered by industry code and employee band
• Regulatory registries and licensing databases that list active firms in each jurisdiction
• Trade association member universes and conference exhibitor lists
• Commercial firmographic databases, cross-checked against the free sources for coverage gaps
• Public filings that reveal installed bases, customer counts or unit volumes

Segment before you multiply

A blended price across enterprise and mid-market customers matches neither. Split the universe by size band or vertical, apply a distinct spend estimate to each, and apply an adoption filter rather than assuming full penetration: the relevant question is how many firms in the universe actually buy products like this, not how many exist. Then validate completeness against a macro anchor, so the count is neither missing whole segments nor double-counting overlapping ones.

Technology install data shows how much this filter matters. One technology-intelligence dataset detects 332,197 companies running Salesforce, a small fraction of the registered companies worldwide. The aspirational universe of 'companies that could use a CRM' and the operational universe of 'companies that demonstrably buy one' differ by orders of magnitude. Install signals, job postings and procurement records shrink an abstract universe into one you can actually size, a discipline that applies equally to VC and PE funds screening targets and to advisors building a market section under deadline, and it is the same evidence discipline that underpins market analysis in due diligence.

When the niche has no published pricing, annual spend per customer has to be reconstructed from evidence. The strongest source is the target's own data room: actual contract values by customer segment, which beat any external proxy. Competitor list pricing, procurement data and management's stated average contract values fill the gaps, and management's numbers should always be tested against booked revenue before they enter the model.

• Work from net, not list price: discounts, channel margins and implementation fees move realized spend per customer materially.
• Separate one-off from recurring revenue, because a market sized on recurring spend behaves differently from one sized on project fees.
• Sanity-check the implied market share: divide the target's revenue by your market estimate and ask whether credible peers have ever attained that share.

The share check catches the most expensive error in niche sizing. In diligence terms, the pattern worth pressure-testing is an addressable market that turns out to be narrower than the plan assumed once growth reaches scale. If your bottom-up market implies the target will capture a share no comparable company has achieved, either the market is overstated or the plan needs explicit justification. Both findings belong in the memo, not in a footnote, which is why financial due diligence and the commercial workstream have to read the same market number.

When a segment has no direct data at all, the practical move is to borrow structure from adjacent markets that share its customers, technology or buying behavior. An adjacency gives you a template for how many customers exist, how they buy and what they pay, which you then adjust for the ways the niche genuinely differs.

• Growth trajectory: is the neighbouring market expanding for reasons that transfer to the niche?
• Margin structure: do comparable vendors earn comparable gross margins?
• Competitive concentration: is share distributed in a way that resembles the niche?
• Capital intensity: does the cost to serve match, or does the niche need a different model?

Unconventional public evidence does much of the work. Job postings reveal how many sellers a competitor employs and which segments they hunt in. Earnings call transcripts disclose revenue per segment for public players. Professional profiles occasionally settle questions outright: one deal team triangulated a niche category's size from a market leader's former head of sales, who disclosed prior-year revenue in a posted resume. None of this appears in an analyst report, and all of it is checkable.

Treat the adjacency-derived figure as a third independent estimate, never a substitute for the build. Comparable-company inference works best as a cross-validation check, because it depends on market share assumptions that are themselves estimates, a limitation that also shapes market analysis for an acquisition.

Reconciliation is where the sizing earns credibility. Set three estimates side by side: the bottom-up build, a narrowed top-down figure that starts from a published aggregate and filters down to your segment, and management's view or the adjacency-derived number. The cross-check is only real if the methods and data sources are independent; running three variants of the same assumption proves nothing.

• Convergence within roughly 15% builds confidence in all three estimates.
• A larger divergence signals a wrong input, most often a hidden double count or an optimistic penetration assumption.
• Test the contested assumption against independent sources, such as competitor revenue disclosures.
• Watch for the classic red flag: a market definition that implies far more revenue than the target actually books.

Finding which input is wrong is the most valuable hour of the analysis. If the bottom-up build and the top-down filter produce dramatically different numbers, one of them rests on a bad assumption, and locating it before the investment committee does is far better than defending it afterwards. Segmented penetration analysis helps: calculating adoption separately by geography, customer size and vertical exposes which sub-segment is carrying an implausible assumption. Materiality-weighted findings and structured red flag reporting keep these gaps visible across the commercial due diligence workstream rather than buried in a spreadsheet tab.

An investment committee does not ask whether the number is right; it asks whether the assumptions survive challenge. Present a range with a base case rather than a point estimate, identify the two or three assumptions that move the result most, and stress-test those with sensitivity swings. Every input should be sourced and dated in the document itself, so the committee can recompute the table rather than take it on faith.

• A recomputable table of segments, customer counts and spend per customer, with a source on every row
• Customer interviews that test the assumed buying behavior, not just the price point
• Internal consistency between market size, the revenue plan and sales capacity

Separate growth from share gain. A plan that quietly requires both an expanding market and rising share is two bets presented as one, and committees price that ambiguity accordingly. The asymmetry also matters: an understated market costs little, while an overstated one costs credibility across the whole memo, including the parts that were right. That is why evidence has to stay source-linked and reusable from screening through investment committee materials, and why the market section of an IC memo should read as a chain of evidence rather than a conclusion. The analysts and partners who defend these numbers face the same scrutiny whether they sit buy-side or advisory.

The method compresses into a sequence that fits a live deal timeline, whether the team sits in M&A advisory firms or on the buy side.

• Define the segment boundary in writing before touching a spreadsheet.
• Ingest the data room and make contracts and financials searchable.
• Extract contract values and customer lists as evidence, not as summaries.
• Build the count-and-spend table by segment, with a source on every row.
• Triangulate against public filings, registries and management data.
• Version the findings with sources attached so the committee can recompute them.

This is the workflow Plausity is built to support: it is the workspace where the evidence is structured, not an automated market-sizing engine. Data Room Ingestion connects to the VDR so contracts and financials become searchable within minutes. The AI-Analysis Engine helps extract and cross-reference pricing and customer evidence across thousands of documents. Risk Radar surfaces materiality-weighted gaps and anomalies in the market case, the Collaboration Hub keeps commercial, financial and legal workstreams aligned, and the Report Builder drafts the IC-ready market section with full source traceability. Built for today's investment and deal teams. Trusted by >200 firms.

For teams going deeper, the guide to market sizing in commercial due diligence covers bottom-up TAM construction in detail, and the commercial due diligence checklist maps the full workstream end to end. The common thread is the same one that runs through this article: a market number is only as strong as the evidence behind each row of the table, and the teams that structure that evidence deliberately are the ones whose numbers survive the committee.

Plausity is an AI-native due diligence and deal intelligence workspace that helps M&A advisory firms, VC and PE funds, corporate development teams and investment-banking teams structure evidence, findings and questions across a data room. Plausity supports evidence extraction, source grounding, findings management and IC preparation — it does not replace human analysts, advisers or investment professionals, does not provide legal, tax, audit, regulatory or investment advice, and does not make autonomous investment decisions. All findings require human review. Built for today's investment and deal teams. Trusted by >200 firms.

To explore the underlying capabilities, see the Plausity AI analysis engine and the findings and risk intelligence product page. For team-level workflows, see how VC and PE funds and M&A advisory firms use Plausity across live deals.

1. wallstreetprep.com
2. prnewswire.com
3. precedenceresearch.com
4. theirstack.com