On Search Leadership

As a high-class consultant, I don’t just talk to companies about the technical aspects of search. I also help them them address organizational challenges.

For example, how should a search team be structured? Should search be combined with recommendations? Should there be a centralized search group, or should each product team manage its own search applications?

One of the biggest organizational challenges companies face is hiring search leadership — specifically engineering and product leaders. It’s hard enough to hire one strong search leader, let alone two. In my experience, companies often find themselves agonizing over which leadership role to prioritize.

Managing search requires both product and engineering strengths.

As I’ve written before, managing search requires both product and engineering strengths. When there are parallel leaders for product and engineering, it’s critical that each be able to lean into the other’s role. And with the right pair of leaders, you can make “two in a box” leadership work.

But it’s pretty hard to find two strong search leaders for engineering and product, let alone two who complement each other and share power well. So, if you can’t hire both immediately, which role should you prioritize?

Search requires a large, complex infrastructure, not just for serving traffic but also for building and maintaining the index. Query understanding, relevance, and ranking are hard science problems that require robust engineering solutions. On top of that, search presents daunting operational challenges. In short, search requires a well-oiled machine; and that requires a strong engineering team that can deliver quality, scale, performance, and reliability.

But you don’t want to build the right solutions for the wrong problems. A strong product leader can prioritize the most important user and business problems, whether that means focusing or particular segments of traffic or making precision-recall trade-offs that reflect real-world costs and benefits. Search involves a collection of problems, and not all of those problems require the same level of quality, scale, performance, or reliability.

A search engineering leader needs enough product sense to make good prioritization decisions. Conversely, a search product leader needs enough understanding of engineering to have reasonable expectations about what kinds of solutions are possible — and practical — to build.

Both matter, but focus on product leadership first.

If you have to decide whether to prioritize engineering or product leadership for search, I don’t envy you. But I’d focus on product leadership first. You need great engineers to build solutions, but start by choosing and framing the right search problems to solve.

And make sure to hire product managers and engineers —not just leaders — who can lean into each other’s roles. Search doesn’t divide nearly into product management and engineering, so try to hire people who can do a bit of both.

ps. Trying to hire a search leader or grow your search team? Check out these interview questions for engineers, data scientists, and product managers.

pps. Where do data scientists fit into all this? Check out my post on where you should put your data scientists, or this great post by Pardis Noorzad.