Ubersuggest / NP Digital · Self-serve SEO

Turning PLG churn into customer lifetime value

A PLG transformation that closed the strategic maturity gap between what the product offered and what SMB customers could actually use.

Halved
Voluntary churn
~3×
Average tenure
2.5×
Paid customer base
Keywords tracked per user
Global
Self-serve scale

Role  Chief Product Officer & Advisor Period  2020–2022 Sector  Self-serve SaaS at global scale

The situation

Demand was not the problem. The platform had enormous self-serve top-of-funnel and a genuinely useful research tool underneath it. But a large share of SMB customers signed up with intent and left before getting anything out of it.

The product worked well for experienced SEO practitioners. It did not work for people who lacked the strategic maturity to turn keyword research into a business result — and that was most of the paying base.

What the real problem was

Customers were choosing keywords based on aspiration rather than competitive reality. They misread domain authority and backlink strength, targeted the wrong geography, and cancelled before the compounding benefit of content had time to appear.

Every exit survey said the product was missing something. It was not. It was failing to walk anyone to a first win. That is a design problem, and it shows up on the income statement as churn.

What I changed

  • Redesigned onboarding from a static setup flow into a guided, visual value exchange — the product asked for less and gave back something usable immediately.
  • Built site scraping and brand inference into onboarding so the product started from the customer’s actual business context instead of a blank field.
  • Improved keyword selection with competitor context, keyword-difficulty guidance, and geolocation targeting for users who could not self-assess winnability.
  • Embedded playbooks for lower-maturity users who needed strategy, not more data.
  • Partnered early with a generative AI provider to build structured long-tail content workflows — moving the product from telling customers what to do into helping them do it.
  • Added reminders, publishing cadence support, and recrawl-based feedback loops so progress became visible before results arrived.
  • Introduced higher-value packaging and laid the groundwork for usage-oriented monetisation as AI compute costs rose.
  • Extended acquisition and activation through a contextual Chrome extension and adjacent brand intelligence.

What happened

Voluntary churn more than halved. Average tenure roughly tripled. Trial-to-paid conversion and ARPU both improved by double digits, and the paid customer base grew by a multiple. Average tracked keywords per customer went up roughly fourfold — the leading indicator that told us onboarding was actually working before the retention numbers moved.

Why this one matters

It is the cleanest demonstration I have that retention is a product design outcome, not a lifecycle-marketing outcome. Nothing here was a discount, a save offer, or a win-back campaign.

Value realisation funnelFive stages from signup to month six, showing the proportion of customers reaching each stage before and after the onboarding redesign.Share of new customers reaching each stageSigns up100 → 100Completes setup62 → 84Tracks 10+ keywords28 → 61Reaches a first win11 → 39Still paying at month six8 → 31BeforeAfter
Nobody was churning over a missing feature. They were churning before the product ever produced a result.

The full version

~17% → 7.6%

Voluntary churn

3.5 → ~10 months

Average tenure

+18%

Trial to paid

+17%

ARPU

~90K → ~240K

Paid customers

Voluntary churn over eight quartersVoluntary churn falling from about seventeen percent to seven point six percent across eight quarters.0%5%10%15%20%Q1Q2Q3Q4Q5Q6Q7Q8
Voluntary churn 17% to 7.6%. No save offers, no discounting, no win-back campaigns.

The numbers

  • Voluntary churn fell from roughly 17% to 7.6%.
  • Average tenure rose from 3.5 months to nearly 10 months.
  • Trial-to-paid conversion improved roughly 18%.
  • ARPU increased roughly 17%.
  • Average tracked keywords per customer went from about 12 to more than 50.
  • Total users grew from roughly 2M to 5M; paid customers from roughly 90,000 to nearly 240,000.

The monetisation detail

We introduced higher-value lifetime plans specifically to improve realised ARPU while tenure was still short — a deliberate bridge, not a strategy. As retention improved, the lifetime plans became a liability against future compute costs, and we built add-ons and tiering to prepare for usage- and value-based pricing before AI inference costs made the old model untenable.

If your PLG business is considering lifetime plans: they work, they buy you time, and you will spend eighteen months unwinding them. Decide knowing that.

What I would do differently

  • Ship the guidance layer before the AI content layer. Customers who could not choose a winnable keyword generated content against the wrong ones faster.
  • Segment onboarding by maturity from the start rather than building one guided path and tuning it. The gap between an agency user and a first-time SMB owner is too wide for one flow.
  • Instrument "first win" explicitly as a metric from day one. We inferred it from tracked-keyword counts for far too long.

← All case studies

Work with me

Something in the business is not working, and the explanations have stopped being convincing.

A first conversation is 45 minutes and costs nothing. If I am not the right person, I will say so and usually know who is.