You Don't Need a Product Manager Until Decisions Start Flooding In

At the very start of a company, hiring a product manager only gets in the way. That is the argument of Saumil Mehta, the founder of the startup LocBox who later led product groups at Square for about ten years, and the timing, he says, depends on how many inputs for decisions are arriving.

He describes the product manager's job as making good choices under uncertainty by gathering signals and fitting them together. On day zero, he says, there are only two signals to fit: the founders' own guesses about the market, and what prospective customers say about their problems. Finding a product that matches the two, what startups call product-market fit, is the founders' job, and handing it to someone else is counterproductive.

That changes once those early guesses start to work. Between roughly the third and sixth month, a startup may add its first engineers, designers, and business hires. By six to nine months, it may have design partners, early customers who agree to test the unfinished product and give feedback. Within 12 to 18 months, inputs are coming from engineers, designers, current and prospective customers, and analytics, the data on usage and website traffic.

Even if the big strategy is settled, that pileup forces fast, tactical choices every day. He lists questions such as "Should you build Feature X or Feature Y first?" as typical of the decisions a team then faces. Should it reshape its plans to please its largest customer, or answer requests from many smaller ones? When is a feature finished enough to release, and is the team moving quickly without giving up too much quality?

Mehta felt that crunch as the unofficial product manager at LocBox. Worrisome churn, customers leaving or canceling, appeared, and the team could not tell if the cause was sales, the type of customer it targeted, account management, or simply that new customers were not using enough features. Its analysis of product engagement was weak because the tracking tools, what engineers call instrumentation, had been deprioritized. Customers were asking for basic features that rivals already offered.

A full-time product manager would have been able to handle that work much faster over a given two-sprint window, or two short work cycles, Mehta says. Instead, he could not get into maker mode, the uninterrupted focus needed for detailed product work, during the few days the team was planning its next sprint. Mehta said of that stretch: "I was helping close a key hire, helping sales improve their prospecting data and looking for a new office space in San Francisco." When your week looks like that, he argues, it is time to hire.