Scoring the TAM: One Rubric for the Whole Market, Not One Deal at a Time

Forwarded to you? I am Heath. I build go-to-market systems and put AI to work in sales, the right way, then I write down exactly what I built, what broke, and what it moved. One story per week, receipts only. This one is about a pipeline full of accounts that all looked the same.

The reflex: sort the pipeline and work the top

Every team with a full pipeline reaches for the same move. The accounts are right there. A thousand of them, sitting in the CRM, and a sales team that has to hit a number this quarter. So you sort by something, last touch, ARR, a stale lead score, whatever is easy, and you tell the reps to work the top.

I did exactly that. I let the reps work whoever was on top of the list, and the list was sorted by nothing that predicted a sale. It failed the same way every quarter. A rep would spend two weeks on an account that was never going to buy, because on the screen it looked identical to the one right next to it that was ready. There was no way to tell them apart, so the whole team guessed, and guessing at scale burns a quarter of your capacity on accounts that were never in play.

The block was never effort. It was that every account looked like every other account. Scoring one deal at a time does not fix that, because the rep only scores the deal already in front of them, and the market is bigger than the pipeline. The signal that separates ready from noise has to run across the whole market before anyone opens an account.

The AE feels this as "which of these thousand do I call first." The CSM feels it as "which of my accounts is actually ready to expand." The marketer feels it as "which segment do I spend the budget on." Same reflex every time: a pile that all looks the same, worked from the top because there is no better order.

The reframe: score the whole market on one rubric

A pipeline is not a priority list. It is a pile until you rank it, and you cannot rank it by scoring the deals one at a time as they walk in. The move is to run one rubric across the entire addressable market, so every account carries a tier and a next move before a human ever touches it.

You cannot prioritize a market by scoring one deal at a time. One rubric across the whole TAM is the priority list.

Same move, other seats. A CSM ranks the book by which accounts show the signals that predict a second team adopting, not by who renews next. A marketer ranks the market by which segment carries fit and intent together, not by who filled out a form. The pile is never the thing you work. The one rubric that ranks it is.

How the best teams frame it

I am not the first person to argue that you score the market, not the deal. The operators who have built account prioritization at scale mostly agree on where the leverage is, and it is not a fancier score on one lead. It is one rubric, run across everything, that ends in a move.

SOURCE

Koen Stam, "The Account Prioritization GTM Playbook"

What it argues. Prioritization is not a one-time list you build and forget. It is a living tier system that combines fit and signal and updates as the market moves, so reps always know which accounts sit at the top and why. The whole point is to work the ranked market top-down, not to react to whatever came in.

My take. Agree, and the tiering is the part teams skip. They stand up a fit score and stop. A score with no tier and no move is just a number, and a rep cannot work a number.

SOURCE

Common Room, "Account Prioritization: A Practical Guide to Signal-Based Targeting"

What it argues. Fit tells you who could buy. Signals tell you who is moving right now. The winning teams combine the two so the account list is ranked by readiness, not just by how well a company matches the ICP on paper. Static firmographics alone leave you working cold.

My take. Extend. Signals earn a place in the score, but only capped. Let intent carry an account on its own and you route a bad-fit account with a loud week. Fit sets the base, signal adds lift, and the cap keeps intent honest.

SOURCE

Demandbase, "Doing B2B Account Scoring the Right Way: Models and Examples"

What it argues. The unit of scoring is the account, not the lead, because buying is a group decision. A good model layers firmographic fit, engagement, and intent into a single account score, then keeps it honest by backtesting against closed-won and re-running it as the data changes.

My take. Agree hard. The reason my first pipeline was useless is that it was sorted at the lead level by nothing that predicted a sale. Score the account, show the work, and let the number carry a tier a rep can act on.

Even a curator has to concede when the field agrees: nobody who has actually built this thinks a better score on one lead is the hard part. One rubric across the whole market, ending in a tier and a move, is the moat.

The method: Solve, Stack, Split

SOLVE THE CRUX

What is the real problem, framed as work and not a headcount?

The problem is not "we need a better account score." It is "every account looks like every other, so nobody can tell ready from noise." So the first work is one rubric that runs across the whole TAM, not the pipeline: ICP fit earns the base, 40 to 70, product engagement adds up to 30, buying signals add up to 20 and are capped. For the AE that rubric ranks who to call first, for the CSM who to expand into next, for the marketer which segment to spend on.

STACK THE CONTEXT

What tech and signals turn a raw account into a tier a human can act on?

Not a shopping trip. Deepline runs the enrichment, scoring, and brief across the whole TAM every week, and holds the ICP fit that earns the base. Amplitude carries product engagement, the loudest and most honest signal, up to +30. Common Room adds the buying signals, capped so they can only help. It all stacks into one number in Deepline and lands in Salesforce, where the rep already works, carrying a tier and a first move.

SPLIT · CUT THE DRAG

What low-judgment work goes to the system?

Enrich, score, stack, tier. Every account in the whole market, every week, with no human in the loop. This is the part that never scaled when a person did it by hand, and the part AI does perfectly because it is the same rubric applied a thousand times without getting tired or playing favorites.

SPLIT · KEEP THE JUDGMENT

What stays human?

The rubric and the call. The operator decides what earns the base, what product signal is worth, and how far intent can lift before it is capped. The rep reads the tier, decides if it is the moment, and owns the conversation that closes. One owner on the rubric, a weekly review of what converted, and the winning patterns feed back into the weights.

The workflow: the board that runs it

Solve, Stack, Split is the shape. Here is the actual board, lane by lane: what the agents run, what stays human, and the tool at each step. Once it is wired, the whole thing runs every week without anyone kicking it off.

The receipt

At a company I was at, a growth-stage B2B SaaS with far more accounts in the market than reps could ever work, this is the exact build I ran. Set one rubric, score the whole TAM on it every week, put a tier and a move on top.

The reps did not work harder. They worked a ranked list top-down, where every account already carried a tier and a first move. We also excluded 12 zero-ARR MQL noise opps so the read stayed clean, because a scoring model is only as honest as what you refuse to let into it. Speed and hit rate came from removing the guesswork, not from adding effort.

Both of my receipts are revenue-side. Drop your own workflow in. A CSM turns the same build into an expansion ranking: fit for the base, product usage for the lift, and a capped intent signal, run across the book so the top of the list is the account most ready to grow. A growth marketer turns it into a segment ranking: the same rubric run across the market to decide where the budget goes, not which form came in last.

WHAT I LEARNED

1. A full pipeline is not a priority list. It is a pile until one rubric ranks it.

2. Score the whole market, not the deal in front of you. The market is bigger than the pipeline, and that is where the order comes from.

3. Cap the intent signal. Fit earns the base, product adds the lift, and intent can only help, or a loud week carries a bad-fit account.

4. A score is not enough. It has to end in a tier and a move, or a rep still has to guess what to do next.

The move this week

Do not build the whole engine. Write down the one rubric, fit for the base, product for the lift, intent capped, and score the next fifty accounts in your market on it by hand. Give each one a tier and a first move. Hand the reps the ranked list and tell them to work it top-down for a week. Watch what converts. That is the model in miniature, and it tells you whether the full build is worth it before you spend a quarter on it.

Two builds that sit next to this one:

• The ICP Playbook — "A firmographic filter is not an ICP; it is a list. Derive the real one from evidence, then your ICP survives contact with a real pipeline." That real ICP is what earns the base of this score.
• The Product Channel — "The question is never whether the channel exists. It is whether your usage data has been turned into context a rep can act on. Map the events first." That product signal is the loudest lift in the rubric.

This is one build from the Build Log. Every week I take one sales or revenue problem, run it through the loop, and show the receipts. If someone forwarded this, the subscribe button is right below. Keep building. Heath.