The company had invested in a self-service platform for SIM card customers and was pursuing a Product-Led Growth motion to drive upgrades and expand usage. The initial approach — feature advertising banners, content blocking, and contact forms to prompt upgrades — was not converting users.

The question was simple on the surface. The answer turned out to be far more complex.

The team believed the problem was visibility. The solution was to make the features more prominent: better banners, stronger calls to action, more aggressive upgrade prompts.

The banners were visually well crafted. They looked good. And users still did not respond.

Rather than optimising the existing approach, I stepped back and asked: who are the customers that actually succeed on this platform, and what does their usage look like?

I evaluated existing customer organisations with significant usage and a significant number of SIMs. I looked at why they were successful, how their platforms were structured, who the actual users were, and how they used the self-service tools day to day.

What I found reframed the entire problem.

The critical insight: these two users have completely different definitions of value — and completely different willingness to pay.

Analysing a cohort of 23 mid-market customer organisations and 71 active platform users surfaced patterns that contradicted the assumption users were passive.

• Nearly half of all SIM update API calls platform-wide originated from mid-market customers in the last twelve months.
• 17 of 23 companies had active API keys making calls at least every six months.
• 8 of 23 were piping webhook events into their own internal systems.
• Around 50% of SIM updates made via the dashboard came from this segment.

Different roles within the same customer organisation used the platform differently. Support engineers lived in the SIMs table and Traffic Monitor. Software engineers opened single SIM pages directly from links in their own tools. CTOs checked network lists and created API keys. CEOs updated SIMs themselves more often than any other role. Hardware engineers opened the specific SIMs that needed attention, bypassing search entirely.

These users were not browsing. They were operating. The interface was a control panel wired into their day-to-day work, not a showcase waiting to be explored.

Successful customers showed a clear pattern. They had created multiple custom network lists. As they expanded geographically or operationally, they created more — specific devices connecting to specific networks in specific business areas.

This was the inflection point. The moment a customer's connectivity needs became complex enough that manual management broke down was the moment the platform became genuinely valuable to them. That was when they were willing to pay for more.

The PLG motion was optimised for the trial user — who rarely reaches that inflection point. It was almost entirely ignoring the scaling user — who does.

The research also revealed a third dimension: who actually uses the platform is not always who decides to buy it.

Designing for one of these people while ignoring the others breaks the entire experience.

There was an additional layer unique to this product: a physical dependency.

When a user purchased a SIM and logged into the platform for the first time, they saw an empty dashboard. Not because something was wrong, but because the physical SIM had not yet activated on the network. The physical world moves slower than the digital one.

The user's emotional state in that moment was confusion and doubt.

The existing response to this was more banners. More information. More content in the empty space.

But the user was not in a reading mindset. They were in an anxious mindset. No amount of beautiful design would fix a mismatch between the user's emotional state and the content being served to them.

Rather than advertising features or blocking content, we explored solutions built around three principles.

The least costly solution — banners with information — had been tried and had failed even when executed beautifully. Users do not read. Users feel things, and they respond to things that meet them where they are emotionally.

The most important question in PLG is not "how do we show users our features" but "who is the user that actually converts, and what does value look like for them at the exact moment they need it?"

The answer to that question changes everything: the empty state, the onboarding, the upgrade prompt, the pricing model, the platform structure itself.

Designing for the user's emotional state rather than the product's feature set is harder to measure and harder to sell internally. But it is the only approach that actually works.