SMART Goal

Writing a goal is easy. What's hard is deciding, when the quarter ends, whether it was achieved. Sentences like "improve customer satisfaction" or "improve code quality" look fine the day they're written and offer nothing to judge against three months later.

Pagedeck ran into this. Last quarter's list had one line reading "improve onboarding." When the quarter closed, Mia and Grace spent twenty minutes on whether to count it as done. The screens had genuinely changed and the copy had been rewritten, so improvement had happened — but the share of free teams that started paying and stayed sat at 4 percent, exactly where it began. Mia judged by the work done. Grace judged by the change produced.

They differed because the basis for judging wasn't inside the sentence. That argument recurs every quarter, and each recurrence makes the next quarter's goals a little vaguer, because a sentence that resists judgment is more comfortable to write than one that invites it.

A SMART goal isn't a new format for writing goals. It's a ruler held up against a sentence already written: run the five tests in order and rewrite whatever fails. "Improve onboarding" fails three of the five, because it says nothing about what changes, by how much, or by when.

Apply the five tests in order.

1. Specific. Check whether the sentence says what changes and for whom. "Improve onboarding" doesn't name which part of onboarding or whose experience. Narrow it: "the steps between signup and the first teammate invitation."
2. Measurable. A number and the place that number gets read from both have to be present. A number without a source means the end of the quarter starts with an argument about which data to look at, and whoever has the more convenient dashboard tends to win it. This test overlaps with the job of a key result — the measurable outcome that confirms whether things actually moved in the intended direction.
3. Achievable. Check whether the height has a basis. Raising conversion from 4 percent to 7 requires being able to say why 7. A calculation like "three drop-off points, roughly a point each" is enough.
4. Relevant. Check whether it connects to what the company is trying to do this quarter. A goal that's well written on its own but disconnected from the direction leaves the company where it was even in a quarter where the goal was met, which is the one failure the other four tests cannot catch.
5. Time-bound. Write the deadline. For a quarterly goal that's the end of the quarter, sometimes with one mid-quarter checkpoint added.

Run all five and "improve onboarding" becomes this: rework the steps between signup and first teammate invitation within three months, raising the share of accounts that invite a teammate inside seven days from 32 percent to 45 percent as measured in product logs. The sentence got longer, and everything that was added is information used to judge it at quarter end. None of the added length is decoration.

Three places suit this well. The first is individual goals: apply the five tests to the two or three items one person owns for a quarter and the call at review time becomes straightforward. The second is execution items, where what to do is already settled and the work is attaching a number and a date. The third is any document several teams read, because a specific sentence tells the team next door what they should be waiting for.

The limits start at the third test. Applied strictly, achievability pulls goals down to a line the team is sure to clear. That runs head-on into a stretch goal — a goal deliberately set at a height that isn't reachable by doing the obvious things, which arrives with an agreement that 70 percent counts as a good outcome. Using both in one document means deciding in advance which items get which tests.

The second limit sits on measurability. Told to attach a number, people choose whatever is easy to count. Documents written and tickets closed are easy to count, and they frequently rise while nothing improves. This test asks for a verifiable change, not for a countable one.

The third limit is about what this is. It isn't a system for managing goals; it's a ruler held against sentences, and a ruler cannot tell you which quarter to hold it up in. Cadence, visibility, and the connection between company and team goals belong to KPIs — the core indicators set once and watched continuously — or to whatever operating method is in use.

"This one fails on measurability. Write down which number, and where we read it."

The sentence carries no basis for judging it, so the argument about it has only been postponed.

"I know the deadline is quarter end, but let's add one checkpoint in the middle."

Splitting the time-bound test into two marks, so a miss shows up while there is still time to act on it.

"We're skipping achievability on this one. It's a stretch item."

An agreement not to apply the five tests mechanically to every item on the list.

After that argument, Pagedeck attached the five tests to the place where quarterly goals get written. In the first quarter of doing so, five of eight items failed the second test — either no number at all, or a number with no stated source.

Fixing them cost one extra meeting. In exchange, the quarter-end retro spent eight minutes on whether the goals were met, because the method for judging was already written into each sentence. The twenty-minute version of that conversation never happened again.

The most common reading is that a SMART goal is a goal-management method like OKRs. Five letters with a name attached look like a system. But this determines neither cadence nor visibility. It only asks whether a sentence is in judgeable form. Running a quarter on OKRs and using these five tests on the key result sentences inside it is closer to how it actually gets used. Treating the two as competitors and asking which to adopt produces no answer.

The other mistake is thinking a good goal passes all five. Try to satisfy every test and what survives is certain, small and safe. In practice teams decide per item which tests apply strongly. Individual execution items get all five; a quarter's stretch items get four, with achievability set aside deliberately and that exception stated out loud. Without that distinction an organization moves, inside six months, toward writing down only what it already knows it can do.