You are in a workshop. The whiteboard has Lead, MQL, SQL, and Opportunity pinned across it. Someone asks: if a contact downloads a whitepaper but their company has eight employees and we sell to enterprises, do they become an MQL?
Half the room says no. They don’t fit the ICP, so they stay in Lead. The other half says yes, let them in, but score them low so sales knows to deprioritize.
Both sides sound reasonable. And the choice you make in that moment will reshape your entire funnel: your conversion rates, your sales team’s trust in marketing, and whether your pipeline report means anything six months from now.
Why this trade-off matters more than it used to
Checkpoint GTM treats this design choice as urgent because the classic MQL is failing: Forrester found the typical lead-centric process converts fewer than 1% of inquiries to closed-won, and proposed scrapping MQLs for an opportunity-centric buying-group model. Where you apply ICP fit, gate or score, decides whether it sharpens or fragments your funnel.
The traditional MQL is under pressure. Forrester’s revenue-process research found that the typical lead-centric process converts fewer than 1% of initial inquiries to closed-won deals, and proposed replacing the entire MQL construct with an opportunity-centric buying-group model. Whether or not you go that far, the stat points to a system-level signal: the way most teams define “qualified” is disconnected from how buyers actually buy.
At the same time, a16z’s Michael King recently published a nine-dimension ICP framework arguing that most founders cannot describe their ICP with sufficient specificity. The downstream cost, he writes, is elevated acquisition cost, low conversion rates, and misaligned product roadmaps. The ICP is the foundation. But where you apply it in the lifecycle, as a gate or as a score, determines whether it sharpens your funnel or fragments it.
The gate: what it does and when it works
A gate means ICP fit is an entry requirement. If the contact’s company does not meet your firmographic criteria: employee count, revenue range, industry, geography, they do not advance to MQL regardless of their behavior.
The upside is purity. Sales only sees contacts that fit. MQL-to-SQL conversion rates tend to be higher because every lead that arrives has already cleared a structural bar. Your pipeline report is cleaner. SDRs spend less time disqualifying.
The downside is volume loss, and invisible volume loss. A gated MQL stage drops contacts silently. Marketing cannot tell if a high-intent, slightly-off-ICP contact is engaging because they are a genuine buyer at a company that just barely misses the threshold, or because they are noise. And the threshold itself is often arbitrary. “Five thousand employees” is a round number. The company just below that line, whose VP is actively requesting a demo, just disappeared from your funnel.
Generally, this approach works best for teams with high inbound volume and limited SDR capacity. If you have more leads than your team can process, the gate is a rational triage mechanism. It does not optimize for precision. It optimizes for capacity.
The score: what it does and when it works
A score means ICP fit is one of several dimensions that determine priority within a stage. The contact advances to MQL based on engagement: a form fill, a demo request, a sequence reply, and then fit scoring determines where they land in the SDR’s queue.
The upside is flexibility. You see everything. The enterprise VP at a company just below the gate threshold is visible. The intern at a startup who downloaded three whitepapers is visible too, but scored lower. Sales can make judgment calls with data instead of flying blind.
The downside is noise. If scoring is not calibrated well, and in most early implementations it is not, sales gets a queue that feels random. The intern and the VP look the same in the list. Trust erodes fast. And once sales stops trusting the score, they stop using the stage entirely and fall back to gut instinct and whatever came in this morning.
Generally, this approach works best for teams in expansion mode: building pipeline across segments, testing new markets, still learning which companies convert. It is also the better fit for account-based motions where multiple contacts at the same company need to be tracked.
Should ICP fit be a gate or a lead score?
Checkpoint GTM tells lifecycle teams to gate when SDR capacity is tight and the ICP is narrow, and to score when running multi-segment or expansion motions. A gate optimizes SDR trust but drops high-intent contacts silently; a score optimizes pipeline visibility but floods the queue with noise. Most teams past Series A need both.
This is really a question about what you optimize for at this stage of your company’s growth.
Gate = optimize for SDR trust. Every lead that hits the queue belongs there. Conversion rates are interpretable. The cost is invisible volume loss.
Score = optimize for pipeline visibility. Nothing hides. The cost is noise, which you have to manage with good scoring design and regular calibration.
Most teams with fewer than three SDRs and a well-defined ICP should gate. Most teams with a multi-segment motion or a sales team that needs to cover mid-market and enterprise simultaneously should score. This needs to be taken with a grain of salt, there are always exceptions, but the pattern holds across the lifecycle workshops we run.
What does gating versus scoring look like in a real funnel?
In the DACH lifecycle redesign Checkpoint GTM ran, the gated enterprise team posted a strong-looking 40% MQL-to-SQL rate but thin pipeline, while the scoring mid-market team converted just 12% yet generated three times the volume, catching fast-growing accounts that bought within 60 days. The fix was running both tracks, not picking one.
We recently worked with a mid-market B2B SaaS company in DACH that was redesigning its lifecycle stages after an acquisition. Two teams, two CRMs, two completely different definitions of “qualified.”
The enterprise team had been gating on ICP fit: minimum employee count, specific industries, budget identified. Their MQL-to-SQL conversion was around 40%, which looked strong on paper. But their total pipeline was thin. Marketing was generating demand that never made it past the gate.
The mid-market team was scoring everything. Every form fill became an MQL. Their conversion rate was 12%, which looked poor. But they were generating three times the pipeline volume and catching accounts that the enterprise team would have filtered out, fast-growing companies that were buying within 60 days.
The fix was not picking one approach for the whole company. The fix was running both: gating for the enterprise segment where SDR time is expensive and the ICP is narrow, scoring for mid-market where speed and coverage matter more. Two lifecycle stage definitions behind one CRM. The automation routes contacts into the right track based on firmographic dimensions, and each track has its own exit criteria.
The five-step design sequence
- Start with your SDR capacity, not your ICP. If your team can only work 50 leads per week and you are generating 200, gate. You do not have a choice. If your team can handle the volume, score.
- Define ICP fit as dimensions, not a binary. Following a16z’s framework, map at least three specific dimensions: company size, industry, and one behavioral or technographic signal. A binary “fits / does not fit” is too coarse for scoring and too brittle for gating.
- Set the gate at the floor, not the ceiling. If you gate, gate on the absolute minimum viable customer: the smallest company, the broadest industry set that could plausibly buy. Everything above the floor enters the stage; scoring sorts from there. Most teams gate too high and lose their fastest-growing segment.
- Calibrate monthly for the first quarter. Pull every MQL from the past 30 days. Check: did the ones that scored high actually convert? Did the ones that scored low actually churn out? If your score is not predicting conversion within one quarter of use, the model needs to be rebuilt, not patched.
- Separate fit scoring from engagement scoring in your reporting. If you blend them into one number, you cannot diagnose whether you have an ICP problem or a nurture problem. Keep fit as one axis and engagement as the other. A high-fit, low-engagement contact tells you something completely different from a low-fit, high-engagement contact, and they need different next actions.
What this means for your next lifecycle workshop
The next time you hit this fork in a whiteboard session, resist the urge to pick one approach for the whole funnel. Ask the room: “Will you market to this segment differently based on this stage?” If the answer is yes, if enterprise and mid-market get different sequences, different SDR playbooks, different demo tracks, then you probably need different stage definitions for each segment. That is not complexity for the sake of it. That is the CRM reflecting how your business actually sells.
Sources
- King, Michael. “A Framework for Defining and Refining Your ICP.” Andreessen Horowitz, May 2025. a16z.com
- Forrester Research. “The Revenue Process Alignment Series, Part 1: The End of MQLs.” Forrester, April 2022. forrester.com
- ChartMogul. “The SaaS Go-To-Market Report.” ChartMogul, Q1 2025. chartmogul.com
Noah Charak