Why consumer pirate metrics break on a sales pipeline
The usual AARRR dashboard follows one person from arrival to payment within a few weeks, and every stage is an event the product itself records. Lead generation stretches and splits that path. The period between first form and signed contract can run to months, so stage counts reported for the same week describe different buyers. The person who filled in the form may never be the one who signs; a finance manager downloads a guide, a director takes the meeting and a procurement team approves the purchase. And half the stages happen in conversations held by sales, which a website analytics tool never sees.
None of this makes the framework useless. Its discipline, one defined event per stage and a rate between each, is exactly what a lead programme often lacks. But the definitions have to be rewritten rather than borrowed, and the comparison of lead generation and demand generation is a reminder that capture is only half of what marketing does here; AARRR as adapted below describes the capture half.
Decisions that change for lead generation
| Stage | Generic AARRR | AARRR for lead generation | Decision it forces |
|---|---|---|---|
| Acquisition | A new visitor or install | A new identifiable contact from a form, event or partner | Which sources count as acquisition and which are only visits |
| Activation | First value inside the product | The lead accepted by sales or sales development | Where the acceptance line sits and who draws it |
| Retention | Coming back to use the product | Leads not ready to buy re-engaging through nurture | What counts as re-engagement and over what window |
| Referral | A user inviting another user | More contacts from the same account entering the pipeline | Whether referral is counted by person or by account |
| Revenue | Payment within weeks | A signed contract a sales cycle later | Which cohort the contract is credited to |
Activation is acceptance, not a click
The product definition of activation, the moment of first value, has no direct equivalent before a sale. The closest analogue is the moment someone with buying-side judgement says the lead is real. In practice that is acceptance by sales or sales development. Set it there, not at a second download or a webinar attendance, because a lead can consume a great deal of content without ever being a prospect. The exact stage names and criteria are defined in the guide to MQL, SQL and lead stages; this page does not restate them. The decision AARRR adds is to treat the acquisition-to-acceptance rate (CVRShare of visits or clicks that convert.) as the activation rate, and to read it by source.
Retention means keeping undecided leads warm
In a long sales cycle, most new leads are not ready on arrival. The retention stage measures whether they come back: a reply, a meeting request or a return visit to pricing within a set window after entering nurture. The work behind it sits in the nurture sequence workflows, and the stage's metric is the share of non-accepted leads that re-engage, not the open rate of the emails. Choose the window from how long buyers typically take to move, and record recycled leads separately so that a lead returned by sales is not counted as fresh acquisition.
Referral happens inside the buying committee
Consumers refer friends; business buyers bring colleagues. When a second or third person from the same organisation enters the pipeline, the account is moving, and deals with several engaged contacts tend to progress differently from those resting on one champion. Count referral as additional contacts per account, from forms, events or introductions by the first contact. Content built for each committee role, described in the guide to business-to-business content, is the main lever. This choice also changes the unit of the whole dashboard: accounts become as important as people.
Revenue is read by cohort of lead creation
Contracts lag leads by a sales cycle, so the revenue row has to be credited back to the week or month the leads were created and read at a fixed age. A weekly AARRR view therefore holds two kinds of row: fast stages read for this week's arrivals, and slow stages read for the cohort that is now old enough to show results. Mark which is which. Cost per customer (CACCost to acquire one new customer.) comes from the same cohort reading, or it divides this quarter's spend by contracts generated by last quarter's.
Split stage owners across three teams
Acquisition and retention sit with marketing; activation sits at the hand-off with sales development; revenue sits with sales. Write the owner next to each stage's KPIA metric tied directly to an objective.. The CRM guide to lead stages and pipeline shows how stage dates are captured, and they are what makes cohort reading possible. Native social forms, explained in the LinkedIn targeting and lead gen forms guide, can inflate acquisition quickly, which is why the owner of activation needs a voice in how acquisition is judged.
Decision points in a worked AARRR for lead generation
An invented payroll-compliance consultancy, Ledgerline Advisory, maps its weekly AARRR view from form fill to signed contract
Ledgerline advises mid-sized employers on payroll compliance. Its sales cycle runs about 12 weeks, so the weekly view reads the fast stages for this week's leads and the slow stages for the cohort created 12 weeks earlier. The table shows the cohort of the week of 6 July, read in early October.
| Stage | Definition used | Owner | Count | Rate |
|---|---|---|---|---|
| Acquisition | New contacts from forms, webinars and partner introductions | Marketing | 240 | – |
| Activation | Accepted by sales development within 10 working days | Sales development | 60 | 25% of acquired |
| Retention | Non-accepted leads who replied, booked or revisited pricing within 60 days | Marketing | 27 | 15% of 180 |
| Referral | Accepted accounts with a second contact added | Marketing and sales | 18 | 30% of 60 |
| Revenue | Signed contracts within 12 weeks of lead creation | Sales | 9 | 15% of accepted |
Average first-year contract value was £8,400, so the cohort produced £75,600 on £14,400 of media spend, or £1,600 per signed client. The constraint sits in an unexpected place. Of the 18 accounts with a second contact, 7 signed; of the 42 with a single contact, only 2 did. Acceptance is not the weak stage; the deals stall when one person carries the case alone. Retention adds a second finding: 6 of the 27 re-engaged leads were accepted on their second pass, a route the old weekly report had counted as new acquisition.
Decisions: marketing produces a short briefing for finance directors that accepted contacts can forward, and tracks second-contact rate as the referral metric; recycled leads get their own row; the weekly report shows this week's acquisition and activation beside the 12-week-old cohort's revenue, never this week's contracts.
Read without the redefinitions, the same quarter would have produced a report that said lead volume was fine and conversion to revenue was poor, with no stage to act on. Defining referral by account exposed the real constraint; defining retention by re-engagement stopped returning leads being double-counted as new; reading revenue by cohort connected contracts to the marketing that actually produced them.
Measuring AARRR for lead generation: KPIs and cadence
Acquisition and activation can be read weekly, by source, for the leads that arrived that week, though acceptance needs a short grace period for sales to act. Retention is read monthly, on the window chosen for re-engagement. Referral, measured as contacts per active account and the share of accounts with more than one contact, moves with the pipeline and is best read monthly. Revenue and cost per signed client belong in a quarterly review, always for the cohort old enough to have closed. The CRM, not the web analytics tool, is the source for every stage after acquisition, so the dashboard is only as reliable as the stage dates sales records. The conversion-rate calculator helps check each stage-to-stage rate.
Where AARRR misleads in lead generation
Volumes are small, so a single large deal or a single bad week swings the lower stages more than any change in marketing. The linear drawing hides that leads move backwards: accepted leads get recycled, opportunities stall and return. A stage owned by sales can change definition without notice when territories or targets change. Counting by person overstates progress in accounts where several people download the same guide without any buying intent. And the framework still starts at the form; the work of creating demand that pays off in later quarters, the territory of the TOFU, MOFU, BOFU framework and the wider FunnelModel of the customer journey stages., never appears as a row.
Common mistakes
- Activation set at a download. Content consumption is not a prospect saying yes.
- This week's contracts beside this week's leads. Two cohorts, one meaningless ratio.
- Recycled leads counted as new. Acquisition rises while the pipeline stands still.
- Referral read as social shares. In a buying committee, the useful referral is a colleague joining.
- No named owner for activation. The hand-off stage drifts between teams and nobody fixes it.
- Judging retention on open rates. Opens measure the email; re-engagement measures the lead.
