A market model you can rebuild in front of the board.

0

TAM, SAM and SOM models built end to end, from company records rather than a vendor slide.

TAM and SAM

The scope of work

Most TAM numbers come from an analyst with a chat window open, and survive exactly one question. Nobody can say which filters produced them, so nobody can defend them when they are challenged.

We build yours from company level records with every filter stored, so it can be rebuilt from scratch, and so it moves when an assumption changes.

How do you build a TAM that survives the first question?

Step 1

We start from who you have already closed

  • Closed won and lost, by segment and size
  • What you charge, and where you can support it
  • Read only. Nothing written, changed or deleted
Who you closed
Who you lost, and why
What you charge them
Where you can support them
The model starts from what actually happened, not from a target.
Step 2

We name every filter

  • Industry, headcount, revenue band
  • Geography you can actually support
  • The tech signals that predict a fit
EVERY FILTER, NAMED AND STORED
Industry and size16 filters
Geography6 regions
Tech signals9 signals
Change one and the whole model moves with it.
Step 3

We build the narrowing

  • Total, then serviceable, then obtainable
  • Each step with a reason beside it
  • Checked against who you have closed
THE NARROWING
TAMevery company that could buy
SAMthat you can reach today
SOMat your win rate
Each step has a reason written next to it.
Step 4

We derive the SOM

  • Win rate by segment and source
  • Real cycle length, not the target
  • What this team can actually cover
WHERE THE NUMBER COMES FROM
Win rateby segment and source
Cyclehow long a deal really takes
Coveragewhat this team can work
The SOM falls out of the modelrather than being chosen first and justified after
Step 5

We hand over the model

  • Every filter stored, so it rebuilds
  • The account list behind the number
  • The assumptions on one page
WHAT YOU GET
The model, with every filter storedrebuildable
The account list behind the numberin your CRM
The assumptions, written downone page
Everything a board member can ask about, answered before they ask.
Step 6

You own the model

  • Built inside the systems you already own
  • Documented so it does not live in one head
  • Rerun it yourself when the market moves
Book time with us →
SapwoodYou
Documented so anyone on your team can rebuild it, not just the person who commissioned it.

The tools we use to build your TAM.

TAM

What usually happens: A number produced by an analyst with a chat window open, or lifted from a vendor slide. Nobody in the room can say where it came from, which filters produced it, or what would change it. It survives exactly one question.

What we build instead: Built from company level records. Every firmographic filter named and stored, so the number can be rebuilt from scratch by someone who was not in the room. If an assumption changes, the model moves with it.

How it is different
Bottom upfrom company records, not an analyst estimate
Every filternamed, stored and reproducible
Rebuildableby someone who was not there
SAM

What usually happens: Usually skipped entirely. TAM gets presented, then SOM gets presented, and the step that explains which of those companies you can actually sell to today never happens.

What we build instead: TAM narrowed by what you can genuinely reach: the segments you have a motion for, the geographies you can support, the tech stacks you integrate with, and the deal sizes your pricing fits.

How it is different
Named constraintsmotion, geography, stack, price
Checkedagainst who you have actually closed
Defensiblewhen someone asks why it is smaller
SOM

What usually happens: Last year plus a growth percentage, presented as analysis. It is a target dressed as a forecast, and everyone in the room knows it.

What we build instead: Derived from your real win rate by segment, your coverage capacity, and how long a deal actually takes. The number falls out of the model rather than being chosen first and justified after.

How it is different
Win rateby segment and source, from your CRM
Capacitywhat this team can actually cover
Derivednot chosen and then justified
Can you rebuild your own TAM from scratch?We will read what you have closed and show you what the model says, free, before you commit to anything.
Get a free audit →

Case study: We did not let ChatGPT draw the territory map.

The territory map
Territories drawn on state lines, because that is how they had always been drawn. A state line tells you where a border is. It does not tell you where the buyers went. Esri and GIS work across four years of customer and demographic movement. The growth was concentrating central and west. So we stopped drawing on state lines, and cut zones at the grain the buyers actually live at.
West
CAORWANVIDUTAZNM
Central
MTWYCONDSDNEKSMNIAMOILWIOKAR
South
TXLAMSALGAFLTNSCNCKY
East
VAWVMDDENJPANYCTRIMAVTNHMEOHINMI
Zone 1Pacific and Mountain WestCA, NV, AZ and the western half of CO
Zone 2Front Range and PlainsEastern CO, UT, WY, NM and into western KS
Zone 3Central corridorMO, AR, TN and the Mississippi river counties
Zone 4Great Lakes and EastUnchanged, because the buyers had not moved
The market, and what we actually took
TAM$8M
Every company that could buy
firmographic pull, filtered and deduped
SAM$4.5M
That we could actually reach
checked against who we had actually closed
SOM$3M
What we predicted we would take
win rate by segment, applied to the zones
Actual$5.5M
Every deal tagged to a zone, product and owner
rolled up and checked years later

Every opportunity carried the zone, product and owner it was predicted under. So the model could be rolled up years later and checked against what really closed. It came in 83% above the SOM, and we could say exactly which zones did it.