The HubSpot Lead Scoring Playbook

Export closed-won and closed-lost deals from the last 12 months, with each associated company. For every one, record its size, industry and region, the title of the person who signed or led it, and the lost reason.

Your closed deals show who actually buys. Set them beside the profile in your strategy deck, and let the backtest in step 06 pick.

"Mid-market and growing" is an opinion. "Employee count between 100 and 500, checked on the LinkedIn company page" is a rule. Write each rule so two people applying it get the same answer, and note how many of your wins match it.

Schools, government bodies and competitors get large negative points, so no amount of engagement lifts them into a top grade. Then list the fit properties that must be filled before a score is trusted.

For each event, count how often it happened in the 90 days before a deal was created, on wins and on losses. An event earns points when it shows up more often before wins.

A pricing page view that appears on 70% of wins and 70% of losses tells you someone was shopping, and little about who buys.

Cap each group so one noisy event cannot fill the score, and set decay to match your sales cycle. HubSpot offers decay intervals of 1, 3, 6 or 12 months.

Watch for the busiest records in the portal. A student, a job seeker or a competitor can out-click a real buyer, and without the disqualifiers they rise to the top.

HubSpot's lead scoring tool sits under Marketing, then Lead Scoring, on Marketing Hub or Sales Hub at Professional or Enterprise. It allows more than one score per object, so the new model runs beside the old one until the backtest reads.

Build it as a combined score. Grades run from A1 to C3. The letter is fit, from A for high fit, and the number is engagement.

HubSpot's test record and distribution preview confirm the arithmetic. Whether the score ranks buyers is the job of step 06.

A1 and A2 go to a rep within a set number of days. A3 is the right company gone quiet, so it goes to nurture and gets reviewed monthly. B1 gets a rep review before any outreach.

C1 is very active at a company that cannot buy, so it gets the resource it asked for and no sales outreach.

HubSpot points cannot tell a blank field apart from a failed rule. Both score zero. So records missing a required fit property go to a review list, where a person looks it up, instead of to a grade.

Watch for a threshold with no action attached. When the customer definition changes, it keeps sorting accounts the old way.

Grade last year's wins and losses and compare the share of each that lands in A. Read the fit letter, since a won customer keeps engaging after the deal and inflates the number on every win.

Sign how sure you need to be before you look. We use a 90% chance that wins really grade A more often.

With 40 wins and 60 losses, 28 of 40 against 24 of 60 is a 99.8% chance, with the real gap somewhere from 13 to 44 points. 24 of 40 against 33 of 60 is 69%, too close to call, so rewrite the fit rules and run it again.

Credit each opportunity to the grade the account held on the first of the month. A meeting raises engagement, so crediting the new grade would let the outcome promote the account after the fact.

Compare A1 and A2 against A3 and B. One month is often too few accounts to call, so pool the months until the chance passes the line.

No difference after three pooled months sends you back to the fit rules. A review list that grows every month means fix the data before the weights.