Lead Scoring That Works: Fit, Timing and Signal Decay

Say your CRM holds 8,000 companies, and your outbound capacity covers about 300 accounts a week. Which 300?

A common answer is one blended score: points for industry, points for company size, points for opening an email, points for a visit to the pricing page. The result is a number nobody can explain. An account that fits perfectly but has nothing happening scores the same as a poor-fit account that clicked 3 links.

This article explains a simpler model we use: 2 separate scores for fit and timing, decay that makes old signals lose value, thresholds that decide who gets what, an example weighting table, and the mistakes that break scoring.

Why One Blended Score Fails

A lead score is a number that ranks accounts by how worth your time they are. The problem with a single blended score is that it mixes 2 different questions.

• Could this account ever buy? That depends on industry, size, business model, region and whether you have won similar companies before. It changes slowly.
• Is now a good time? That depends on what changed recently: a new leader, a funding round, a job post for the role your product supports. It changes every week.

When both go into one number, a high score can mean "perfect fit, nothing happening" or "weak fit, lots happening", and the rep cannot tell which. They stop trusting the score and go back to picking accounts by gut feel.

Splitting the score fixes this. Each account gets a fit score and a timing score, and the action depends on the pair.

Fit Score vs Timing Score

Plot the 2 scores against each other and every account lands in 1 of 4 places.

Fig. 1Fit vs timing

Fit score (low to high)

Watch

Right account, no reason yet. Monitor for signals.

Act now

Right account, live signal. Research and reach out this week.

Ignore

Wrong account, no signal. Leave it off the list.

Check fit

Live signal, weak fit. Review by hand before any outreach.

Timing score (low to high)

Fit score comes from your ideal customer profile. If you have built an ICP matrix (1 ICP split into 3 segments by 2 personas), fit measures how closely an account matches one of those segments. Our guide How to Build an ICP Matrix covers that step.

Timing score comes from signals. Each signal adds points when it appears and loses points as it ages. 15 B2B Buying Signals, Ranked by How Fast They Go Stale lists the signals we watch and how long each stays useful.

Fit acts as a gate. An account with no fit should never reach a rep, however many signals it shows. Timing decides the order among accounts that pass the gate.

Example Weights

Here is a weighting table to start from. These weights are an example, not a benchmark. Your own numbers should come from looking at which accounts became meetings and deals.

In this example, fit tops out at 100. Timing can exceed 100 if several signals stack, which is fine: an account with 3 fresh signals deserves to rank above one with 1.

A few principles behind the weights:

• Firmographics carry fit. Industry, size and revenue are facts about the company, and they change slowly.
• Signals carry timing. Behavior like email opens does not belong in either score. Opens are unreliable and say little about buying intent.
• Fewer factors beat more. Every factor needs data to feed it. A factor you cannot fill for most accounts adds noise.

Signal Decay

Decay is what keeps a timing score honest. Without it, an account that raised money 14 months ago still looks hot, and reps waste the week on stale reasons.

Here is how one signal might decay over time in our example.

Fig. 2How a signal decays (example)

1. Week 0 to 2
   Full points

   20 points. The buyer is forming plans and choosing vendors.

2. Week 3 to 6
   Most points

   14 points. Still a strong reason to reach out.

3. Week 7 to 12
   Half points

   10 points. Plans are set; the window is narrowing.

4. After 12 weeks
   Expired

   0 points. Wait for a new signal.

Different signals decay at different speeds:

• Fast decay (4 to 6 weeks): content engagement, job posts. These reflect a short window of attention or need.
• Medium decay (8 weeks): tech stack changes. A team that just adopted a tool is open to related changes for a while.
• Slow decay (12 weeks): new leaders, funding, expansion. These start planning cycles that run for a quarter.

To make decay work, every signal needs a date stored on the record, and the timing score needs to recalculate on a schedule, daily is enough. The signals in CRM with decay play shows how to store signals so this happens automatically.

Thresholds and Routing

Scores only matter if they change who does what. Thresholds turn the 2 scores into 4 actions, and routing rules decide who owns each.

Fig. 3From score to owner

1. 01Recalculate daily

   Fit stays stable; timing updates as signals arrive and age.

2. 02Apply thresholds

   Each account lands in act now, watch, check fit or ignore.

3. 03Route act now

   Researched brief and draft go to the account owner for approval.

4. 04Route check fit

   A person reviews by hand before anything is written.

5. 05Review the model

   Quarterly, compare scores against meetings and deals won.

Here is an example set of thresholds, using the weights above:

• Act now: fit 70 or higher and timing 30 or higher. The account is researched, a brief and draft go to the owner for approval, and outreach starts this week.
• Watch: fit 70 or higher, timing under 30. No outreach yet. The account stays on a watch list and moves to act now the moment a signal pushes timing over the line.
• Check fit: fit between 40 and 69, timing 30 or higher. A person looks at the account by hand. Sometimes the fit data is wrong; sometimes the account is a genuine edge case worth a try.
• Ignore: fit under 40. The account leaves the outbound list, whatever its timing.

Routing follows ownership. If an account already has an owner in your CRM, it goes to them. If it is an existing customer or has an open deal, it goes to the account manager or deal owner, never into a cold sequence. The HubSpot account tiering and scoring play shows one way to build tiers and routing inside HubSpot.

Keep the act now list sized to your capacity. If your team can work 300 accounts a week and 900 qualify, raise the timing threshold until the list fits. A list larger than your capacity means the best accounts wait in a queue while their signals decay.

A worked example

Here are 3 hypothetical accounts scored with the example weights and thresholds above.

Account A: a B2B software company, 120 staff, in a target segment. Fit scores 90: industry, size, revenue, model and region all match. Last week it hired a new VP of Sales (20 points) and posted 3 SDR roles (15 points). Timing is 35. Result: act now. The account owner gets a brief that leads with the new leader and the open roles, plus a draft first email for approval.

Account B: the same profile, nothing happening. Fit is 90, timing is 0. Result: watch. No outreach this week. If it announces funding next month, timing jumps to 20; if it also posts a relevant role, timing reaches 35 and it moves to act now automatically.

Account C: a consumer brand that engaged with your posts. It liked 3 posts and visited your site (10 points of timing), but sells to consumers and sits outside every segment. Fit is 25. Result: ignore for outbound, however active it looks. Engagement from a non-fit account is pleasant and rarely turns into a deal.

Now imagine Account A 4 months later with no new signals. The VP of Sales hire has fully decayed and the job posts have expired. Timing is back to 0, and the account moves to watch. If nobody replied during the live window, the system waits for the next reason rather than repeating the old one.

Where the Scoring Data Comes From

Fit and timing are only as good as the fields beneath them.

• Fit data comes from firmographic sources: company databases such as Apollo, your own CRM fields, and the company website. Fill industry, headcount and region for every account before scoring; a blank field should score 0, not an average.
• Timing data comes from signal sources watched on a schedule: job boards, funding news, LinkedIn for leadership changes, tech stack detection, and your own content engagement. Each signal writes a dated record to the account.
• Ownership and status come from your CRM: current customer, open deal, assigned owner. These override scores in routing.

Tools like Clay or n8n can pull and combine these sources, and HubSpot or Salesforce can hold the scores as properties that recalculate on a schedule. Keep the score and its parts visible on the account record, so a rep can see that an account scored 35 on timing because of a new leader and 2 job posts, and judge for themselves whether that is a good reason to write.

How to Know the Model Works

A scoring model is a guess until you test it against outcomes. Every quarter, look at accounts scored in the previous quarter and ask:

• Did act now accounts book meetings at a higher rate than watch accounts? If not, the timing signals or thresholds need work.
• Did meetings from high-fit accounts turn into qualified opportunities more often? If not, the fit factors are wrong.
• Which signals preceded won deals? Raise their weight. Signals that never precede meetings should lose weight or leave the model.

Report this alongside your weekly numbers. The Weekly Pipeline Dashboard covers the 9 numbers we track, including meetings booked by segment, which is where scoring problems show up first.

Common Mistakes

These are the patterns that break scoring:

• One blended score. Mixing fit and timing hides why an account ranks where it does.
• No decay. Old signals keep accounts looking hot long after the window closed.
• Scoring email opens. Privacy features load tracking images automatically, so opens add noise, not intent.
• Too many factors. A 30-factor model is impossible to explain and usually rests on fields that are half empty.
• Scoring dirty data. Duplicate accounts and wrong industries produce confident wrong scores. Clean the CRM first.
• Never reviewing the weights. A model built once and never checked against won deals drifts away from reality.
• Ignoring existing relationships. A customer or open deal should never be routed into a cold sequence because it scored well.

FAQs

Build Your Scoring Model With Oriel

Oriel is the engine room behind the go-to-market team at B2B companies above $100k/month in revenue. We build fit and timing scores inside your own CRM, store every signal with a date so timing decays on its own, set thresholds that match your team's capacity, and route act now accounts to a researched brief and a draft that a person approves before anything sends.

On a 30-minute call, we will look at how you pick accounts today and sketch the first version of a fit and timing model for your market. If you are not ready for this yet, we will tell you that too.