Calibrating a Growth Target Before It Gets Gamed

Marketing Academy · Field Work●Mental Models

All projects·From: Goodhart's Law: When Metrics Become Targets

MiniReverse-Engineer· 25 minutes

Lenskart

Objective: Given a proposed 'double MQLs by Q4' target with 6 months of historical MQL-to-SQL conversion data, calibrate a paired target that survives being gamed instead of shipping the raw volume number.

You're the demand-gen lead at Lenskart and the CMO wants to double marketing-qualified leads (MQLs) by Q4, from 240/month to 480/month. You have six months of historical MQL-to-SQL conversion data to work from.

Model what happens to conversion rate if MQL volume doubles at the current qualification bar, then propose a paired target with a guardrail that keeps quality from collapsing.

Before you start

What you'll need

Free path (everything below is enough to finish)

Google Sheets

FreeModel the current funnel, the gaming-risk scenario, and the paired-target proposal

Free, and a funnel model at this scale needs nothing more than formulas

The process

2 steps

Step 01 of 02

Modeling the gaming risk of a raw volume target

The lesson's real example: a team told to double SQLs hit the number by loosening qualification, and close rate fell from 22% to 11%, with net revenue actually dropping.

Given the historical data below (240 MQLs/month, 35% MQL-to-SQL rate, 20% SQL-to-close rate), what is the most likely way the team hits 480 MQLs/month if the qualification bar is the only lever available, and what happens to downstream conversion if they use it?

Google Sheets— Open 'MQL Historical Data' sheet, review the last 6 months.

Procedure

Record the current funnel: 240 MQLs -> 84 SQLs (35%) -> 17 closed-won (20% of SQL)
Identify the fastest lever to double MQL count: lowering the lead-score threshold that defines 'qualified'
Estimate the funnel if MQL count doubles via a lowered bar but downstream conversion rates fall proportionally to lead quality (assume MQL-to-SQL rate roughly halves to 18%)

Sample output

CURRENT FUNNEL (monthly)
  240 MQL -> 84 SQL (35%) -> 17 closed-won (20% of SQL)

IF MQL BAR IS LOWERED TO HIT 480/MONTH
  480 MQL -> ~86 SQL (18%, lower-quality leads convert worse) -> ~14 closed-won (16% of SQL, sales trusts these leads less)

RESULT: MQL count doubles, but closed-won deals FALL from 17 to ~14/month despite 'hitting' the target.

Healthy

The model is built and shown to the CMO before the target ships, changing the conversation from 'can we hit 480' to 'what target actually grows revenue.'

Unhealthy

The 480 target ships as-is, and six months later someone asks why 'MQLs doubled but the pipeline looks the same.'

What this means

A volume target with no downstream guardrail predictably gets hit by lowering quality, this is modelable in advance, not a surprise after the fact.

So what do I do about it?

Symptom	Action	Effort
A leadership target is set as a raw volume number with no mention of the conversion funnel below it	Model the funnel impact of the easiest gaming move before the target is finalized, not after it's missed	30 min

YouYou can do this yourself, no engineering access required.

Step 02 of 02

Designing a paired target that moves closer to the real business outcome

Rule 2 of the lesson's playbook: move the target closer to the real outcome you have data for. Revenue is harder to game than MQL count.

Propose a revised Q4 target for the CMO that still shows growth ambition but is paired with a guardrail on closed-won deals or close rate, so hitting the number can't happen by lowering quality alone.

Google Sheets— Same sheet, new tab 'Revised Q4 Target Proposal'.

Procedure

Propose an MQL growth number that is ambitious but realistic (e.g. +50% instead of +100%) at the current qualification bar
Pair it with a guardrail: 'MQL-to-SQL conversion rate cannot fall below 30% (down from 35%, allowing some natural variance)'
State the projected closed-won number under the paired target, and compare it to the un-paired 480-MQL scenario from Step 1

Sample output

REVISED PROPOSAL
  Target: 360 MQLs/month (+50%, not +100%), at the current qualification bar
  Guardrail: MQL-to-SQL rate must stay at or above 30% (baseline 35%)
  Projected: 360 MQL -> ~113 SQL (31%) -> ~23 closed-won (20%)

COMPARISON
  Raw 480-MQL target (ungated): ~14 closed-won/month
  Paired 360-MQL target (guardrailed): ~23 closed-won/month

The smaller, guardrailed target produces MORE actual revenue-generating deals.

Healthy

The CMO signs off on the paired 360-MQL/30%-guardrail target after seeing the closed-won comparison table.

Unhealthy

The CMO keeps the 480 number because it 'sounds more ambitious' without seeing the downstream deal-count comparison.

What this means

A target's ambition should be judged by its effect on the real outcome, not by the size of the headline number.

So what do I do about it?

Symptom	Action	Effort
A leadership target gets set purely by how big the percentage increase sounds	Always attach a projected real-outcome number (deals, revenue) next to any proxy-metric target before it ships	30 min

YouYou can do this yourself, no engineering access required.

Final deliverable

A one-page target proposal memo: current funnel, gaming-risk model, and a revised paired target with a projected closed-won comparison.

See a reference example

Sample output

Zendesk, support-ticket-deflection target proposal (excerpt)

CURRENT: 1,200 tickets/month, 42% self-service deflection rate
RAW TARGET REJECTED: '60% deflection by Q4' modeled to encourage agents hiding the 'contact us' link, projected CSAT drop of 6 points
PAIRED TARGET PROPOSED: '50% deflection, guardrail: CSAT cannot fall below 4.2/5' -> projected deflection gain achieved without the CSAT risk

Success criteria

You're done when you can:

Funnel model shows the specific gaming move and its downstream effect on closed-won count
Paired target includes a numeric guardrail, not just 'watch quality closely'
Final comparison shows the guardrailed target produces a better real-outcome number than the raw target