What Sage and Echo-U taught me about the cost of a feature

The support desk is your best research instrument

Working around contact-centre operations gives you an unfiltered view of a product that no analytics stack provides. Analytics tells you where users stopped. A support queue tells you what they believed was going to happen and why they were wrong, in their own words, at volume, timestamped.

Most companies treat contact reasons as a cost line to reduce. It is far more valuable as a product backlog with the priority already calculated for you: the top five contact drivers, ranked by volume and handling time, are almost always a more honest roadmap than the one in the planning tool.

The habit I have kept is simple. Before any roadmap discussion, read fifty recent support conversations. It takes an hour and it has changed my mind more often than any dashboard.

Churn is decided in the first thirty days

Renewal conversations feel like where retention is won. They are not. By the time a renewal is on the table, the outcome was largely determined by whether the customer reached a moment of real value in the first month and whether anyone noticed if they did not.

This is why activation instrumentation matters more than almost any other measurement in early SaaS. If you cannot answer 'what percentage of accounts reached first value within seven days', you cannot manage churn, you can only be surprised by it.

At Wocal that number was concrete: had a venue published its first live listing? Everything in onboarding was designed to shorten the distance to that single event, and it did more for retention than any feature we built in the same period.

Buyer, user and payer are three different people

In consumer products these roles collapse into one person. In enterprise SaaS they almost never do. The person who signs is optimising for risk and price, the person who uses it is optimising for time saved, and the person whose budget it comes from is optimising for a line in a spreadsheet they have to defend.

Product teams that design only for the user build software that is loved and not renewed. Teams that design only for the buyer build software that is bought and not used, which is worse, because it produces the quietest and most complete form of churn.

The practical move is to make the product produce evidence for the payer as a by-product of the user getting value. Usage reporting, time-saved summaries and audit trails are not features for the user, they are ammunition for whoever has to justify the renewal.

Pricing is a product decision

Enterprise SaaS taught me that packaging changes behaviour more reliably than features do. Where a capability sits in your tiering determines who adopts it, how heavily it is used, and whether a heavy user is profitable. That is product design, whatever the org chart says.

For AI-enabled products this has become urgent rather than academic. When usage carries a real marginal cost, an unlimited plan is a promise to lose money on your most engaged customers. Packaging is now the main defence of gross margin.

What I do differently because of it

• Read the support queue before the roadmap, every time.
• Instrument one activation event and defend it above all other metrics.
• Design reporting that lets the payer justify the spend without asking you for it.
• Treat pricing and packaging as part of the product brief, not a follow-up task.
• Estimate the support tail of a feature before committing to build it.

Where this fits in the framework

These lessons live in the Launch and First Revenue stages of The Crossley Method, which is the point where most founder-built products discover that the commercial model and the product design were never reconciled.