Andrew Crossley
    UK · AI SaaS SELF-SERVE

    Product-Led Growth Consultant

    A product-led growth consultant designs the self-serve mechanics that let an AI SaaS product sell itself: onboarding that gets a user to value in minutes, a free tier or trial that converts without a sales call, activation metrics that actually predict retention, and pricing that matches how the product creates value. Engagements start at £4,000 for a focused PLG audit, or run as a fractional retainer of £4,000-£10,000 a month for ongoing build and iteration.

    WHO THIS IS FOR

    You are probably reading this because one of these is true

    • Your signup-to-activation numbers are dismal and nobody can say exactly where users drop off.
    • You're running a sales-led motion for a product that's simple enough to sell itself, and it's burning cash and cycle time.
    • Your free tier either gives away too much or gates the exact feature that would prove your AI actually works.
    • You have activation metrics but they're vanity numbers — logins, not the moment a user got real value.
    • Your pricing was set once, early, and nobody has revisited it against actual usage patterns.
    • You want product-qualified leads flowing to sales instead of every trial being treated the same.

    THE FRAMEWORK

    The Crossley Method: idea to first revenue in seven stages

    See the full method
    1. STAGE 1DiscoverWeek 1
    2. STAGE 2ValidateWeek 2
    3. STAGE 3PrototypeWeek 3
    4. STAGE 4Build MVPWeeks 3-4
    5. STAGE 5LaunchWeek 5
    6. STAGE 6First RevenueWeek 6
    7. STAGE 7ScaleOngoing

    WHAT THE WORK COVERS

    The mechanics of a PLG motion, in order

    PLG isn't a single feature, it's a chain: get someone in fast, get them to value faster, measure the moment that predicts they'll stay, and let usage tell you who's ready to pay more. Most AI SaaS products break somewhere in this chain.

    Onboarding & time-to-value

    Cutting signup friction and getting a new user to a genuine 'aha' moment — ideally in one session, not after a week of setup and an onboarding email sequence.

    Free tier & activation

    Designing what's free, what's gated, and what activation event actually correlates with retention, rather than measuring signups or logins as a proxy for success.

    Pricing & packaging

    Aligning price to the unit of value the product delivers — seats, usage, outcomes — so upgrading feels like a natural consequence of getting value, not a sales negotiation.

    PQLs & expansion

    Instrumenting usage signals that flag when an account is ready for a sales conversation, an upsell, or an expansion nudge, so sales spends time where it converts.

    HOW IT RUNS

    How a PLG engagement runs

    1. Week 0

      Intro call, free

      We look at your signup funnel and current metrics together. If the real blocker is distribution rather than product mechanics, I'll tell you that plainly — PLG can't fix a traffic problem.

    2. Weeks 1-2

      Funnel and activation audit

      Full walk-through of signup to first value as a new user would experience it, plus your existing analytics, to find the actual drop-off points and the metric that should be your activation signal.

    3. Week 3

      PLG plan

      A prioritised set of changes to onboarding, free tier limits, pricing and instrumentation, ranked by expected impact on activation and conversion, not by internal opinion.

    4. Weeks 4-10

      Build and test

      Working with your team or your AI tooling to ship onboarding changes, gating adjustments and PQL scoring, then reading the resulting funnel data before iterating again.

    5. Ongoing

      Iterate on real usage

      PLG is never 'done' — the highest-leverage version of this work is a recurring retainer where activation and expansion metrics are reviewed monthly and the motion is adjusted accordingly.

    PRICING

    Pricing, published

    PLG work splits naturally into a diagnostic phase and an ongoing iteration phase. Most AI SaaS companies need both, but the diagnostic alone is often enough to unblock the next six months.

    PLG audit

    £4,000

    Two-week fixed scope. Funnel walkthrough, activation metric definition, and a prioritised plan for onboarding, free tier and pricing changes.

    One day a week

    £4,000-£6,000/mo

    Ongoing PLG ownership: reviewing activation data, iterating onboarding, and refining the free-tier and pricing model as usage patterns emerge.

    MVP build for self-serve flows

    £12,000-£25,000

    A six-week build for teams that need the actual onboarding, gating and instrumentation shipped, not just designed — sized to the scope of the self-serve flow.

    Building rather than hiring? Get a number in 60 seconds with the MVP cost calculator.

    COMPARISON

    Product-led vs sales-led growth

    PLG isn't automatically the right answer. Here's the honest comparison for an early-stage AI SaaS product deciding which motion to invest in.

    FactorProduct-ledSales-led
    Cost to acquire a customerLow per customer once the funnel works, scales without headcountHigh — every deal needs a rep's time, scales with sales headcount
    Time to first revenueCan be minutes to days if onboarding is goodWeeks to months per deal, gated by sales cycle length
    Best fit productClear, fast value; low price point; individual or small-team buyerComplex value proposition; high price point; committee-based buying decision
    Founder effort requiredHeavy upfront investment in product and instrumentationHeavy upfront investment in a sales process and hiring
    Data you getRich usage data on exactly how customers get valueRich qualitative data from direct customer conversations
    Enterprise dealsStruggles alone — usually needs a sales-assist layer for larger accountsBuilt for this — handles procurement, security review, custom terms
    Common AI SaaS patternPLG for individual/team tier, sales-led for enterprise — a hybridPure sales-led often over-invested in for products simple enough to try alone

    PROOF

    The work behind the advice

    Wocal — £2.7M pre-money

    Founder and Chief Product Officer, 2020-2025. Built the onboarding, pricing and activation model that took a hospitality SaaS product from a single venue to 300+, largely without a traditional outbound sales team.

    Co-Ride — since Nov 2025

    Founder and Chief Product Officer of a community carpooling platform, currently designing the self-serve onboarding and activation loop from first principles at pre-seed.

    Just Eat, 2021-2025

    Exposure to marketplace growth mechanics at scale, including how activation and engagement metrics are used to predict retention long before revenue outcomes are visible.

    Sage and Echo-U

    Enterprise SaaS and contact-centre experience that shaped a clear view of where sales-led motion is genuinely necessary versus where it's just organisational habit.

    Why most AI SaaS onboarding fails before activation

    AI products have a specific onboarding problem: the value depends on the model doing something impressive with the user's own data or context, which usually requires setup — connecting a data source, uploading a document, configuring a workflow — before the 'aha' moment can happen. Every one of those setup steps is a place a curious but unconvinced user drops off.

    The fix isn't removing the AI, it's finding a smaller, faster path to a genuine demonstration of value using minimal setup: a sample dataset, a pre-filled example, a single-click starting point that shows what the product does before asking the user to trust it with anything real. Time-to-value should be measured in minutes for a self-serve AI product, and if it isn't, that's the first thing to fix, before touching pricing or the free tier.

    Most teams instrument this badly. They track signups and logins, not the specific event that represents 'this user has seen the product work'. Getting that single metric right — and it is usually one specific event, not a composite score — is the single highest-leverage piece of PLG work for an early AI SaaS company.

    Designing a free tier that proves the AI works

    Free tiers for AI products get this wrong in one of two directions. Either they gate the exact feature that demonstrates the AI's value, leaving free users unable to tell if the product is any good, or they give away enough that there's no reason to upgrade, especially once inference costs are involved. Both mistakes come from designing the free tier around cost containment rather than around what needs to be proven to a new user.

    A better approach: give unlimited or generous access to the core value moment, and gate on dimensions that scale with genuine usage or team growth — volume, data connections, seats, advanced configuration. This lets a user fully experience why the product works before they hit a wall, and the wall they hit is naturally tied to the value they're already getting, which makes the upgrade decision easy rather than adversarial.

    For AI products specifically, inference cost makes 'unlimited' free usage genuinely expensive, which is why usage-based gating — a monthly credit or query allowance — tends to work better than all-or-nothing feature gating. It protects margin while still letting the product prove itself fully within the free allocation.

    Pricing and packaging that match how the product creates value

    The most common pricing mistake in early AI SaaS is charging per seat for a product where value scales with usage, output or outcomes rather than headcount. If one power user gets ten times the value of a casual one, seat-based pricing leaves money on the table and creates weird incentives — teams sharing logins to avoid extra seats, for instance.

    The alternative isn't necessarily pure usage-based pricing either, which can create unpredictable bills that scare off self-serve buyers who want to know what they're committing to. A hybrid — a base fee for access plus usage-based components at the margin — tends to work best for AI products: predictable enough to buy without a call, scalable enough to capture value from heavy users.

    Packaging should also map cleanly onto the free-to-paid funnel already discussed. If the free tier proves the value and the gate is a natural usage ceiling, the paid tiers should remove that ceiling in clear, understandable steps, not introduce a completely different mental model at the point of payment.

    PQLs: getting sales involved without becoming sales-led

    • A product-qualified lead is an account whose usage pattern indicates they're likely to convert or expand, not just anyone who signed up for a trial.
    • Good PQL signals are specific to your product: number of active users on an account, a particular feature used repeatedly, or usage approaching a plan limit — not generic engagement scores.
    • Sales should only be looped in when a PQL signal fires, for high-value or usage-capped accounts. Most self-serve conversions should never see a human at all.
    • This hybrid model — self-serve for individuals and small teams, sales-assisted for accounts showing enterprise-scale usage — is the realistic PLG pattern for most B2B AI SaaS products, not a pure self-serve-only motion.
    • Getting this wrong in either direction is expensive: too much sales involvement kills the efficiency PLG is meant to deliver; too little means genuinely large accounts churn without ever getting the attention that would have expanded them.

    FREQUENTLY ASKED

    Product-led growth consultant questions

    What does a product-led growth consultant actually do?
    They redesign the mechanics that let a product convert and expand without a sales-led process: onboarding and time-to-value, free tier and gating design, pricing and packaging, and the activation and PQL metrics that tell you which accounts are ready to buy or expand.
    Is PLG right for every AI SaaS product?
    No. PLG suits products with fast, clear value and an individual or small-team buyer who can say yes without procurement. Complex, high-price, committee-purchased products usually need a sales-led or hybrid motion, and a good consultant will tell you which one fits before starting work.
    How much does a PLG engagement cost?
    A fixed-scope PLG audit is typically £4,000 and takes two weeks. Ongoing iteration runs £4,000-£6,000 a month as a fractional retainer, and a full onboarding or self-serve flow build is £12,000-£25,000 depending on scope.
    What's the difference between activation and engagement metrics?
    Activation is the specific moment a new user experiences real value for the first time, usually a single well-defined event. Engagement metrics like logins or session count are proxies that often don't predict retention nearly as well and can mislead teams into optimising the wrong thing.
    How do you price an AI product with variable inference costs?
    Usually a hybrid: a base subscription for access plus usage-based pricing at the margin for heavy consumption, which protects margin on expensive inference while keeping the entry price predictable enough for self-serve buyers.
    What is a PQL and how is it different from an MQL?
    A product-qualified lead is identified by in-product usage behaviour that predicts conversion or expansion readiness, rather than a marketing-qualified lead, which is identified by content engagement or firmographic fit before the person has used the product at all.
    Can PLG and sales-led motion coexist?
    Yes, and for most B2B AI SaaS products this hybrid is the realistic model: self-serve for individuals and small teams, with sales looped in only when usage signals indicate an account is enterprise-scale or ready to expand.
    How quickly will I see results from PLG changes?
    Onboarding and activation changes typically show measurable funnel movement within two to four weeks of shipping, since the sample size for a self-serve funnel accumulates fast. Pricing and packaging changes take longer to read clearly, usually a full billing cycle or two.

    GET IN TOUCH

    Tell me where your funnel is leaking

    A short note about your product and current signup-to-activation numbers is enough. I reply within 48 hours and I'll tell you honestly if PLG is the right lever to pull.

    Your details are used only to reply to you and are never sold or shared. See the privacy policy.