Score, Headline, Issue, Remedy: Anatomy of a GazeSite Finding

Most feedback about websites is useless, and it’s useless in one of two opposite ways. Either it’s a number with nothing behind it, or it’s an essay with no number in front of it. “Your site scores 61” tells you that something is wrong but not what. “Your hero section fails to communicate value to first-time visitors because the messaging assumes prior context” tells you what someone thinks is wrong but not how much it matters, or what to do at nine on a Monday morning. I spent a long time consuming both kinds before noticing that the feedback I actually acted on always had the same shape. It had four parts: a score, a headline, an issue, and a remedy. When I built GazeSite, I made every finding take exactly that shape, and I want to explain why, because the shape is useful whether or not you ever use my tool.

Start with the score. A score is the most abused element of feedback, because it’s the easiest to produce and the easiest to worship. On its own it’s nearly meaningless. What a score is actually for is triage. When a review of your site surfaces fifteen findings, you cannot treat them equally, and you shouldn’t. The score’s job is to make the ordering of your attention obvious: fix the 3s before you polish the 8s. That’s all. A score is a sorting key, not a verdict. The moment you start optimizing the number instead of the site, the number stops measuring anything.

Then the headline. This sounds like a cosmetic detail, and it’s the part people most often skip. It isn’t cosmetic. Forcing a finding into one short line is a test the finding either passes or fails. “The primary call to action is invisible on mobile” is a real finding. If a problem can’t be stated in a headline, one of two things is true: the reviewer doesn’t understand it yet, or it’s actually several problems wearing a trenchcoat. Either way, the compression failure is information. There’s a second function too. A page of headlines can be scanned in thirty seconds, and scanning is how busy people decide what to read. Feedback that can’t be scanned doesn’t get read, and feedback that doesn’t get read might as well not exist.

The issue is where the evidence lives. A headline makes a claim; the issue backs it up. What exactly is on the page, where, and why it hurts. This is the part that earns trust, because it’s checkable. When a reviewer writes that the body text renders at twelve pixels over a busy background image, you can look at the page and see whether that’s true. Specificity is what separates observation from opinion. “The design feels dated” is an opinion; nobody can act on a feeling. “The page uses five typefaces and none of them establish hierarchy” is an observation; you can count. I’ve come to use checkability as the acid test of the issue section. If you can’t verify it by looking, it isn’t a finding yet.

And then the remedy, which is where nearly all feedback in the world quietly gives up. Describing a problem is much easier than prescribing its fix, which is exactly why the prescription is where the value is. A remedy has to be concrete enough to hand to someone. “Improve your messaging” is not a remedy. “Replace the headline with a plain statement of what the product does, in the vocabulary a first-time visitor would use” is. The test I apply is this: could a competent person who has never seen the site before read the remedy and know what to do first? If not, the finding is a complaint with good posture.

The remedy also does something subtler: it keeps the reviewer honest. It’s easy to sound wise pointing out flaws. Committing to a fix means the criticism can be wrong in public. If the prescribed change gets made and nothing improves, the finding was bad, and now everyone knows. Criticism without a remedy is unfalsifiable, and unfalsifiable criticism accumulates like sediment. This is, incidentally, why so much feedback stays vague. Vagueness is safe.

The four parts check each other, and that’s the real design. The score without the issue is a mood. The issue without the remedy is a complaint. The remedy without the headline is advice you’ll never find again when you need it. Assembled, they form something closer to a unit of work than a piece of commentary: this matters this much, here’s what it is, here’s the evidence, here’s what to do. You could drop it straight into a task tracker, which is precisely the point.

In GazeSite, every reviewer’s output is forced through this template, one finding at a time, across each of the six areas it grades. But I’d make the stronger claim that the shape is worth adopting anywhere feedback happens. Next time someone reviews your work, or you review someone else’s, try holding the feedback to the four-part standard. How bad, in one line, what exactly, and what now. You’ll find that most feedback fails the standard. You’ll also find that the small fraction that passes is worth more than all the rest combined. Good feedback isn’t a talent. It’s a shape, and shapes can be learned.