In a set of experiments run by four economists and published in the American Economic Review, when a group needed someone to take on a task that helped everyone and advanced no one, women volunteered far more readily than men, and managers handed out the same tasks unevenly too. The researchers, Linda Babcock, Maria Recalde, Lise Vesterlund, and Laurie Weingart, had a name for that kind of task: low in promotability. Useful to the organization. Invisible on a performance review. The kind of work that holds a team together and quietly stalls the person doing it.
Read the product manager job description again with that phrase in mind. Almost all of it qualifies.
The PM job is mostly glue
Tanya Reilly, an engineer who has thought harder about this than most, calls it glue work: noticing when someone is blocked and unblocking them, writing the document that gets everyone aligned, onboarding the new hire, chasing the dependency nobody else wants to own, turning a messy meeting into a decision. Her point was aimed at engineers who drift away from code and toward coordination. But a product manager does not drift toward glue. Glue is the job. Strip the coordination, the alignment, the unblocking, and the writing out of a PM role and there is almost nothing left to measure.
That is the trap, and it is structural, not a matter of effort. The better a PM is at the work, the less visible the work becomes. A launch that goes smoothly looks like it was always going to go smoothly. The three weeks you spent reconciling two teams who wanted opposite things show up in the record as “no issues.” Nobody files a ticket for the escalation that did not happen because you saw it coming. Your best quarters are the ones where the least appears to have gone wrong, and “nothing went wrong” is the hardest thing in the world to put on a promotion packet.
Why the glue disappears when someone writes your review
Reilly tells a story that should make every PM wince. An engineer spends two years doing the connective work a team depends on, improving process, mentoring, keeping projects from falling apart. Her reviews are glowing the whole time. Then she goes up for promotion and is turned down, because the committee looks for technical contribution and she has shipped very little code. The glue held everything together and left no artifact with her name on it.
Promotion committees reward things they can point to. A shipped feature. A system someone designed. A number that moved and a person who can be cleanly credited for moving it. Glue work resists all three. It is distributed across other people’s wins, it rarely produces a standalone object, and its value is counterfactual: it shows up as the disasters that never arrived. When the person deciding your level scans for evidence and finds a trail of other people’s successes that you happened to be near, the easy and wrong conclusion is that you were support staff for the real work.
PMs are especially exposed here because the role already lives in the gap between credit and responsibility. Engineering ships the code. Design owns the pixels. Sales closes the deal. The PM owns the outcome and almost none of the artifacts, which is a fine way to be accountable for everything and creditable for nothing. I wrote recently about how honest feedback thins out as a PM gets more senior. The glue-work trap is the same problem wearing different clothes: the more your contribution lives in the spaces between other people’s work, the harder it is for anyone, including you, to see and name it.
Who ends up holding the glue
The economists did not stop at defining the problem; they measured who pays for it. In their field data, after an email went out asking for a committee volunteer, 7 percent of female faculty raised their hand against 2.6 percent of the men. Scaled up across a career, the authors of the related book The No Club estimate the median woman spends about 200 more hours a year on non-promotable work than the median man, roughly a full working month. The same research found women are 48 percent more likely to volunteer for such tasks and that managers of both genders are 44 percent more likely to assign them to a woman.
That tilt matters on PM teams, which skew toward exactly the collaborative, facilitative reputation that attracts the “can you just handle this” request. If you are the one the room turns to when something needs smoothing over, that reputation is an asset to everyone around you and a tax on your own trajectory. It is worth knowing whether the glue on your team is being shared or quietly accumulating on the people least likely to say no.
Making invisible work legible
I spent two decades in IT operations, a discipline that is almost pure glue. Operations is the function that is noticed only when it fails; a perfect year looks, from the outside, like a year in which operations did nothing. I watched genuinely excellent people stall out because their value was entirely counterfactual, the outages that never happened, and the review process had no column for that. What eventually moved careers was not doing more glue. It was making the existing glue legible.
A few things worked, and they translate directly to product work.
Keep a record the review process can read. Not a feelings journal, a ledger of decisions and their consequences. The integration risk you flagged in March that would have slipped the launch. The two teams you got to a shared spec in a week instead of a month. The feature you talked the room out of before it burned a quarter. Counterfactual value needs a narrator, because it leaves no natural trace, and if you do not narrate it, the silence gets read as absence.
Attach the glue to an outcome, not an activity. “I ran the cross-team sync” is an activity and reads as busywork. “I owned the dependency between platform and billing that was the single biggest launch risk, and it shipped on the original date” is the same glue, framed as the result it produced. Reilly’s advice to glue workers is blunt and correct: get the title that legitimizes the work, create artifacts that show impact, and have the direct conversation with your manager about what actually earns a promotion here rather than assuming the good reviews will compound. If the honest answer is that the glue will never count, her last resort is the uncomfortable one, which is to stop doing it for a while and go produce something the committee can point to.
And if you manage PMs, the fix is mostly yours. The research is clear that glue accretes on the same people unless someone distributes it deliberately. Rotate the onboarding, the release coordination, the thankless cross-team wrangling. Give public credit for the save that prevented the disaster, not just the ship that made the demo. Decide whether your ladder actually rewards the work your team depends on, and if it does not, stop being surprised when your most reliable people leave to go somewhere their contribution has a name.
The uncomfortable truth underneath all of this is that being indispensable to a team and being promotable are not the same thing, and they sometimes pull in opposite directions. The work that makes you the person everyone relies on is often the work that makes you easy to leave exactly where you are. Worth knowing which kind of work you spent this quarter doing, before someone else decides what it was worth.