Company Priority

Inside a single team, arguments about order don't last long. The same people are looking at the same work, so they agree quickly on what is more urgent. The trouble starts at the boundary between teams. A feature sales asked for and a cleanup engineering has been deferring share no common scale of urgency. One side is talking about a contract closing this month and the other about an outage six months out, so even sitting in the same room neither convinces the other.

When that happens, the order still gets set, just by other means. The team that pushes hardest goes first, or whichever customer complained most recently moves to the front. Pagedeck went through a stretch where two teams had booked the same engineer for the same week and every meeting turned into "ours first, please." Nothing had been decided, yet everyone believed their own work was number one, and the result was three projects each about half finished.

Company priority is the device that lifts this judgment to one place. Work that teams can't compare against each other only acquires an order when someone above both teams builds a single list. Leaving that seat empty doesn't remove the order — it just sets it in a way nobody is accountable for.

It works by collecting the candidates into one list, scoring them against the same criteria, then drawing a line where the resources run out and deciding that everything below it doesn't happen.

1. Collect the candidates in one place. Scattered across team documents, the comparison never starts. Sales requests, product work, and paying down technical debt all go on the same table even though their natures differ. Technical debt is the accumulated cost of cleanup deferred to move faster, and if it never reaches the table it gets deferred indefinitely.
2. Pick two or three criteria and score only against those. Size of revenue impact, number of customers affected, and how much the risk grows if you wait are usually enough. Stretch the list to seven criteria and every project scores high on at least one, which makes all of them important again. Holding the same three criteria across every candidate is what makes a sales request and a cleanup comparable at all: one touches $230,000 of revenue and no other customer, the other touches no revenue and every engineer's working day.
3. Write the resource ceiling as a number. Something like "twelve engineers, three projects running at once this quarter." Without that number the list simply keeps growing.
4. Draw the line and publish what falls below it. This step is the heart of company priority. Naming items one, two and three is easier than writing down that items four and below will not happen this quarter, which is why the second half usually gets skipped.
5. Keep the decision with one person. The only people who can compare two teams' work are the ones accountable for both. Try to settle it by consensus and everyone concedes one item to everyone else until the list is back up to eight.

Write down eight projects while holding resources for three, and the remaining five aren't a list of things you decided against. They're a list of things that will be half-attempted and left unfinished. That is where numbering a list and actually prioritizing part ways.

The most common venue is quarterly planning. It's the meeting where goals get reset every three months, and when each team brings what it wants to do, the total always exceeds what the company can staff. Building the order effectively becomes that meeting's conclusion.

The resulting order gets used again on urgent requests that arrive mid-quarter. When a new request comes in, the question isn't whether it's a good idea but whether it ranks above the current number three. If it does, the team decides together what comes down; if not, it becomes a candidate for next quarter. Without that test, every urgent request adds one more slot.

The same list carries into attaching people and budget. With no order set, resources go out in the sequence requests arrived. With an order set, they attach from the top down until they run out.

Finally, it gets announced at the all-hands, the meeting where the whole company gathers to hear where things stand. Everyone hears what the quarter will and won't include in the same words. If six or seven people out of forty-five sit in the planning meeting, this is the only channel the rest have.

"That isn't in the company priorities this quarter. To add it, the payments rework at number one has to come down."

Reframing a request as a swap rather than an addition.

"We staffed down to number three. Four and below aren't happening this quarter."

Stating the line explicitly so expectations get set in advance.

"Company priorities haven't been set yet, so each team is going on its own read."

Reporting that while the decision waits, the order has scattered into individual judgment.

Three things came to Pagedeck's last quarterly planning meeting. Sofia brought a permissions feature demanded by a customer on a $230,000 contract, Theo brought the deploy pipeline that takes forty minutes to run, and Mia brought onboarding changes aimed at first-week drop-off. All three had a case, and all three together needed more than twelve engineers.

Grace took the first two and pushed onboarding to next quarter. Not because it mattered less, but because running three at once would leave all three half done. The item that came down wasn't deleted; it was written in as the leading candidate for next quarter, which is what let Mia stop working on it this quarter.

The most common mistake is assuming a numbered list means priorities are set. A document running from one to eight looks like finished work. But if all eight have people attached, that document is a list rather than an order. What tells you the decision actually happened is the point where resources stop. With no cut-off, every team proceeds with its own item and the quarter ends with eight things around 60 percent complete.

Reading it as an urgency ranking is almost as common. Urgent work is visible and loud, so it naturally moves forward. But put whatever a customer shouted about yesterday at number one every time, and work that pays off in six months never starts. The pattern is easy to spot at year end: the same platform cleanup shows up in four consecutive quarterly plans and never once gets staffed. It works better to reserve separate capacity for urgent work and build the order out of importance.