08.03
Count by cohort or by calendar week, and stick to it
Understand the two ways to count a week, when each one misleads you, and pick one for your scorecard.
Video walkthrough
- Calendar week: everything that happened this week
- Cohort: what happened to this week's leads, however long it takes
- A worked example with the same leads counted both ways
- Which one to use, and how to label it
What you're building
A decision, written on your scorecard, about how you count, plus the habit that makes it work: either a simple calendar-week count, or a cohort row per week that you update as those leads move forward.
Why it matters
Real estate is slow. A lead from this week may book next week and sign in two months. If you divide this week's signings by this week's leads, you are dividing people who are not the same people. Neither method is wrong; mixing them is.
What you need
- Your scorecard from chapter 08.01
- Two or three weeks of past numbers, if you have them
Time and cost
- Time
- 20 minutes
- Cost
- $0
Step by step
Read the two methods side by side
Calendar week Cohort Question it answers How busy were we this week? How well did this week's leads convert? Held appointments counted Every meeting held this week, from any week's leads Meetings held by leads who arrived this week, whenever they happen Can a rate pass 100%? Yes No When is the row final? Sunday night Weeks or months later (you keep updating it) Effort Low Higher: you revisit old rows Check: You can say in one sentence what each method answers.
Work through the example
A fictional team gets 10 leads in week 1 and 10 in week 2. Week 1 leads book 4 appointments, all held in week 2. Week 2 leads book 3 appointments, held in week 3.
Same activity, two counts (fictional) Row Leads Booked Held Booked to held Calendar week 1 10 4 0 0% Calendar week 2 10 3 4 133% Cohort week 1 (final) 10 4 4 100% Cohort week 2 (final) 10 3 3 100% Calendar counting makes week 1 look like a disaster and week 2 look impossible. Cohort counting shows both weeks held every appointment, but you only know that after week 3.
Check: You can explain why calendar week 2 shows 133%.
Choose your method
- Start with calendar week if you are new to a scorecard. It is simple and shows activity. Judge rates on 4-week totals, which smooths the timing effect.
- Move to cohort when you want to compare sources or campaigns fairly ("did October's Meta leads convert better than September's?"). Tag each lead with its arrival week and count stages per tag.
Check: Your scorecard's definitions block names the method, and the tool's counting switch matches it.
If you choose cohort, tag the arrival week
Add a tag or field like
cohort-2026-W45when each lead arrives (a workflow step in HighLevel, or a tag in the Follow Up Boss event or automation). Each Monday, update the counts on every cohort row from the last 12 weeks by counting stages per tag.Check: Filtering your CRM by one cohort tag returns exactly that week's new leads.
What success looks like
- One method is chosen and written on the scorecard.
- Nobody compares a calendar-week rate to a cohort rate.
- If using cohort, every lead carries its arrival-week tag.
Copyable templates
cohort-YYYY-Www (ISO week, e.g. cohort-2026-W45 for the week starting Monday 2 November 2026)Try it
Weekly scorecard
Your numbers stay in this browser only. Targets are yours: nothing is compared to an outside benchmark.
Your targets (optional)
| Rate | This week | Target | Flag |
|---|---|---|---|
| Lead to qualified 0 of 0 | n/a | - | Not enough data |
| Qualified to booked 0 of 0 | n/a | - | Not enough data |
| Booked to held (show rate) 0 of 0 | n/a | - | Not enough data |
| Held to signed 0 of 0 | n/a | - | Not enough data |
| Lead to signed (overall) 0 of 0 | n/a | - | Not enough data |
| Cost | This week | Limit | Flag |
|---|---|---|---|
| Cost per lead | n/a | - | Not enough data |
| Cost per qualified conversation | n/a | - | Not enough data |
| Cost per booked appointment | n/a | - | Not enough data |
| Cost per held appointment | n/a | - | Not enough data |
| Cost per signed client | n/a | - | Not enough data |
No target missed (you have not set any targets yet). Record the week and compare it with the last 4 weeks.
"Small base" means one person moves that rate by more than 10 points. Judge it on 4-week totals.
Switch between Cohort and Calendar week with booked 9 and held 12: the tool explains why the warning differs.
How to test it
Take one past month. Count booked to held by calendar week and by cohort. If the four-week totals differ a lot, your bookings are spread across weeks and calendar-week rates will be noisy; judge them only on rolling totals.
Troubleshooting
Cohort rows keep changing and the team finds it confusing.
Show cohort rows only once they are 4 or more weeks old, and use calendar week for the current month.
Leads have no arrival-week tag.
Use the contact's created date instead: filter by date created for the week and count stages for those contacts.
Advanced technical details
Optional. You can follow every step above without these.
Technical detailsLag and censoring
Recent cohorts are incomplete ("right-censored"): their leads have not had time to convert, so their rates look low. Compare cohorts at the same age, for example "held within 21 days of arrival", rather than comparing a 2-week-old cohort with a 10-week-old one.
Sources
A method with illustrative numbers. Replace the examples with your own. Last checked 2026-10-20.
- Funnel vs. Stages Distribution Widgets in HighLevelHighLevel Help Center · read 2026-10-20