Customer Access Is Now Production Infrastructure

2 November 2026
Today's argument
As building becomes faster and features become easier to copy, direct and repeatable customer access becomes a core delivery capability rather than a separate research activity.
The cost of producing software is falling, but the cost of being wrong is not.
Teams can now create prototypes, add features and test technical approaches faster than before. Competitors can reproduce visible functionality quickly. At the same time, companies are reconsidering what a business application should do, where AI belongs and whether customers trust the result. An assistant can look useful in a demo and still fail as a product. A new feature can ship smoothly and still give customers another reason to leave.
In that environment, I believe customer access should be treated as production infrastructure.
Most companies do not operate that way. Customer contact is usually fragmented across sales calls, support tickets, research projects and account reviews. Product managers may receive summaries. Engineers may see selected issues. Designers may speak to customers during discovery, then lose access once delivery starts. Churned customers often disappear from view entirely.
That arrangement was already weak when building was slow. It becomes dangerous when building speeds up. A team that can implement an idea in days can also create waste in days. Faster execution only helps when the team can distinguish a valuable customer problem from an internally attractive solution.
Customer access is not the same as collecting more feedback. A large feedback board can create the illusion of evidence while removing the circumstances that gave each request meaning. “Add an export” could mean a customer needs to satisfy an auditor, move data to a competitor or repair a broken internal process. Those are three different product problems.
Access means the team has a reliable way to speak with relevant customers, observe their work and return after release. It includes recruitment, consent, scheduling and a record of what was learned. It also includes customers who stopped using the product, not only the cooperative users who remain engaged.
I would make this a shared capability rather than the private network of a researcher or product manager. Product, design and engineering should all participate. Customer success and sales should enable access, but not act as permanent translators. When every conversation is filtered through the commercial relationship, teams learn what the account wants to negotiate rather than how the product actually fits into the work.
This does require discipline. Direct contact can produce anecdote-driven roadmaps if the loudest customer gets treated as the market. The answer is not to keep teams away from customers. The answer is to combine conversations with usage patterns, commercial context and a clear target segment. A customer conversation should improve a decision, not automatically create a feature.
I also think product leaders should inspect customer access with the same seriousness as technical dependencies. If a roadmap assumes that a team will validate a new workflow with finance managers, who can recruit those people? How long does it usually take? Can the team contact customers without asking three executives for permission? Will they be able to observe the workflow after launch? If the answers are vague, the roadmap contains an unpriced dependency.
This matters for developing people as well. Junior product managers and engineers do not build judgment by receiving polished conclusions from senior colleagues. They build it by seeing contradictions: what customers request versus what they use, what buyers value versus what operators tolerate, and what looks simple in a meeting versus what survives daily work. Removing that exposure may make the team look efficient while weakening its future decisions.
Features are becoming easier to produce and easier to copy. Customer understanding is harder to copy because it depends on repeated contact, accumulated context and the ability to change your mind. I would invest accordingly. Before asking whether a team can build faster, I want to know whether it can reach the right customers without friction and learn from them without hiding behind a dashboard.
This is an automatically generated daily column written in my own voice. The news sources I follow only serve as inspiration for what is topical — nothing here is a summary of, or a quote from, any single article.
Inspired by what was in the air at: lennysnewsletter.com, techcrunch.com, tpgblog.com, mindtheproduct.com, romanpichler.com
- customer-research
- product-operations
- product-leadership
- decision-making