How to calculate TAM: top-down vs bottom-up

Guide · Market Growth · 7 min read · last verified 2026-07-24

The short answer: two roads to the same number

There are two ways to calculate TAM (total addressable market), and a credible number uses both. Top-down starts from a published category size and multiplies by the slice you can serve: category revenue × your addressable share. Bottom-up starts from the ground up: number of addressable accounts × annual contract value (ACV). Top-down tells you whether the market is big enough to matter; bottom-up tells you whether your specific product and price can reach it. When the two land in the same order of magnitude, you have a defensible TAM. When they diverge by 10×, one of your assumptions is wrong — and finding out which is the entire point of doing both.

Top-down TAM: start from the category, cut it to your slice

(Every market figure in this section is a cited third-party analyst estimate, with sources named inline. The share percentages applied to them are a worked example and labelled hypothesis.)

Top-down works when analysts already publish a number for your category. The method is subtractive — you take a big reported figure and remove the parts you cannot serve:

• Find the category revenue. Use the analyst report closest to what you actually sell. For AI-answer optimization, GEO services were ~$1.01B in 2025, projected to $17.0B by 2034 at a 45.5% CAGR [IntelMarketResearch]; a second house puts 2025 at $848M growing to $19.8B by 2034 [MarketIntelo]. That the two disagree is a feature, not a flaw — note it.
• Add adjacent budgets you can capture. Magrios also touches competitive-intelligence software (~$0.6–5B in 2025 depending on how narrowly you scope it, ~15% CAGR toward ~$15B by 2033 [SkyQuest, Fortune Business Insights, Mordor]) and the AI-measurement slice of SEO software (~$41–86B category in 2025 [Grand View, Precedence]).
• Apply your addressable share. You cannot sell to the whole category. As an illustrative cut, if you can realistically address roughly 5–10% share of that adjacent SEO budget plus the full GEO/AEO layer, the constructed top-down TAM lands near ~$8–12B in 2025 (derived), carried toward ~$30–40B by 2030 by GEO/AEO's 34–50% growth.

The strength of top-down is speed and third-party credibility. Its weakness is that it inherits every definitional argument baked into the source report.

Bottom-up TAM: count the accounts, multiply by ACV

(Figures in this section are sourced inline — category numbers attributed to the named analysts, and Magrios’s own numbers labelled derived or hypothetical worked examples.)

(The account counts below are cited; the ACV and the resulting TAM are Magrios's own reasoning, labelled hypothesis — not a market claim.)

Bottom-up ignores the category headline and rebuilds the market from unit economics. It is the more honest of the two because every input is something you can defend from your own funnel:

• Count addressable accounts. For Magrios, realistic buyers are mid-market and enterprise B2B firms with genuine marketing budgets. In the US alone there are ~220,000 companies above $10M revenue, of which fewer than 18,000 are enterprise [SalesHive, Martal, Close].
• Scale to your geography. Applying a ~3.5× global multiplier (US ≈ 28% share of the global mid-market-plus economy — hypothesis) and adding ~120,000 funded startups and agencies gives ~890,000 addressable B2B organizations globally.
• Multiply by ACV. Using a base blended ACV of ~$22K/year (derived from real Magrios pricing — Pro at $750/user/mo, Pro+ at $2,097/user/mo), the bottom-up TAM is ~890,000 × $22,000 ≈ $19.7B (hypothesis).

Bottom-up forces you to name your ACV and your account universe out loud — which is exactly why founders resist it. It is much harder to hide an optimistic assumption in an account count than in a category headline.

Top-down vs bottom-up: a side-by-side comparison

(Figures in this section are sourced inline — category numbers attributed to the named analysts, and Magrios’s own numbers labelled derived or hypothetical worked examples.)

| Dimension | Top-down | Bottom-up |

|---|---|---|

| Starting point | Published category revenue | Count of addressable accounts |

| Formula | Category size × addressable share | Accounts × ACV |

| Best when | Analysts already size your category | You know your ACV and buyer profile |

| Main strength | Fast, third-party credible | Grounded in your real unit economics |

| Main failure mode | False precision; inherits definition fights | Optimistic account counts and ACV |

| Confidence label | Measured category, derived share | Hypothesis (your assumptions) |

| Magrios example | ~$8–12B constructed from GEO/AEO + CI + SEO slice | ~$19.7B from ~890K orgs × ~$22K ACV |

When each method misleads you

(Figures in this section are sourced inline — category numbers attributed to the named analysts, and Magrios’s own numbers labelled derived or hypothetical worked examples.)

Each method fails in a predictable direction, and knowing the direction is how you catch the lie.

Top-down inflates through borrowed precision. Analyst reports disagree wildly for young categories — GEO estimates span from $7.3B by 2031 [Valuates] to ~$19.8B by 2034 [MarketIntelo], a gap driven purely by what each house counts as "GEO." If you quote a single figure to two decimal places, you are inventing precision the underlying data does not have. Top-down also tempts you to claim a share of a category you do not actually serve — the classic "1% share of a $50B market" slide that means nothing.

Bottom-up inflates through optimistic inputs. The two levers — account count and ACV — both bend upward under pressure. Counting every company with a website as "addressable" is how a real ~150K-account market becomes a fictional 5-million-account one. Overstating ACV does the same quietly: the difference between a ~$13K conservative and ~$34K aggressive blended ACV (hypothesis) changes the same account base's TAM by more than 2.5×. Name your ACV assumption explicitly, or the number is unfalsifiable.

How to reconcile the two: triangulation

(Figures in this section are sourced inline — category numbers attributed to the named analysts, and Magrios’s own numbers labelled derived or hypothetical worked examples.)

(The reconciled figures below combine cited category data with Magrios's own hypothesis inputs; both are labelled.)

Triangulation is the practice of running both methods and trusting the answer only where they overlap. For Magrios, top-down lands near ~$8–12B and bottom-up near ~$19.7B — different numbers, but the same order of magnitude. The honest conclusion is therefore a range: TAM ≈ $10–20B and rising (derived), not a false-precision point estimate.

When the two methods agree in magnitude, publish the range and move on. When they diverge by 10× or more, do not average them — debug them:

• Reconcile the definitions. Is top-down counting a broader category than bottom-up's buyer profile? Usually yes.
• Check the account count. Bottom-up too high often means "addressable" was defined too loosely.
• Check the ACV. Bottom-up too low often means you priced the whole market at your cheapest plan.
• Publish the range, label the confidence. A market you can only size to ±one order of magnitude is a market where anyone quoting a single number is guessing — and saying so out loud is a credibility signal, not a weakness.

This is the discipline Magrios applies to its own category and to customers': sizing markets from public, cited evidence, showing ranges instead of false precision, and labelling every forward-looking or bottom-up figure as hypothesis. The GEO/AEO market is the live example — a category analysts disagree on by more than 10×, precisely because it is new.

A worked example, end to end

Here is the full loop as an illustrative worked example, using a hypothetical vertical-SaaS vendor. Suppose analysts size its category at $4B (measured). Top-down: if the vendor can serve 25% share of that category, top-down TAM = $4B × 0.25 = $1B. Bottom-up: the vendor counts 40,000 addressable accounts and charges a $20K ACV, so bottom-up TAM = 40,000 × $20,000 = $800M. The two land within roughly a fifth of each other, so the founder can defend a TAM of ~$0.8–1B (derived range) — and, critically, can point to which number moves if the addressable share or the ACV assumption changes. That auditability is the deliverable, not the single figure.

Where TAM stops and SAM/SOM begin

(Figures in this section are sourced inline — category numbers attributed to the named analysts, and Magrios’s own numbers labelled derived or hypothetical worked examples.)

TAM is only the ceiling. Once you have a defensible TAM, subtract what you cannot reach today to get SAM (serviceable addressable market), then what you can realistically win in three years to get SOM (serviceable obtainable market). For Magrios, SAM is the marketing-mature, English-first B2B core — ~150K–250K organizations × ~$18–22K ACV ≈ $3–5B (hypothesis) — and SOM is the honest small slice a pre-launch company can actually capture. The method is the same at every layer: two independent calculations, reconciled, with the confidence labelled. If you only ever produce one number, you have produced a guess with a decimal point.