One Queue, And A Verify Button That Means It
Every recommendation from every screen, in one list, ordered by priority band. Mark something done and the next analysis goes and checks, with the method recorded and the answer allowed to be no.
Free plan: one website of up to 900 pages, and three re-analyses. No card, no trial clock.
- 5
- States that survive a re-run
- 4
- Verification outcomes
- 0
- Queue rows stored
- Free
- Cost
The Work Gets Done And Nobody Can Prove It
SEO work has a verification problem. Somebody fixes forty canonicals, ticks the row, and nothing ever confirms it. Three months later the same finding is back and nobody can say whether it was never fixed, fixed and regressed, or fixed on a template that got redeployed.
The other failure is subtler: a queue projected from stored descriptions drifts away from the screens it came from, until the task list and the audit disagree about what the site looks like.
- A spreadsheet of SEO tasks that stopped matching reality two months ago
- Fixes marked done with nothing that ever checked
- Regressions found by accident, a quarter after they landed
- A queue mixing observations you cannot finish with work you can
State Is Stored. Everything Else Is Projected.
The queue is projected from the current audit and the current insight run and never stored, so it cannot drift away from the screens it draws on. The only thing stored is what a person decided: recommended, in progress, completed, verified, dismissed. Those states survive every re-run, because they are keyed on the rule or insight subject rather than on a row id.
Lanes say which part of the website somebody is being asked to think about: audit, search, content, setup. Natures say whether there is work attached: issue, opportunity, observation. Observations are shown apart from the queue, because a queue whose items cannot be finished is a queue people stop trusting.
Verify goes and checks, with a method chosen per rule and the reason recorded for each: re-read the affected pages and run the rule over fresh markup, re-fetch the file the site publishes once, or say plainly that only a full analysis can settle it. A rule missing from that map falls back to the conservative answer, because a confident Verified badge on a check that never ran is worse than no badge at all.
The queue
- Every recommendation from the audit and the insight engine in one list
- Grouped by priority band, then impact, effort and reach
- Lanes: audit, search, content, setup
- Natures: issue, opportunity, observation, with observations kept out of the queue
- The audit's score and the insight engine's score are never mixed in one sort
Verification
- Re-read the affected pages and run the rule over fresh markup
- Re-fetch the file the site publishes once
- Or say that only a full analysis can settle it
- Outcomes: passed, partial, failed or unavailable
- A rule with no method mapped falls back to the conservative answer
Settling, on every analysis
- Work found fixed is verified
- Confirmed fixes that came back are reopened
- Work marked done but still detected gets a record saying so
- A check that did not run is never read as fixed
What You Get, And What It Will Not Do
What you get
- A task list that cannot drift away from the audit it came from
- Fixes confirmed by re-reading the page, not by somebody ticking a box
- Regressions caught by the next analysis rather than by accident
- Priority that survives being re-run next month
What it will not do
- It will not claim a fix it could not check. Where only a full analysis can settle a rule, it says so instead of showing a badge.
- Observations are deliberately kept out of the queue, because items that cannot be completed erode a queue people rely on.
- It does not assign work, set due dates or integrate with a task tracker.
About Action Center
What happens to my progress when the site is re-analysed?
It survives. State is the only thing stored, keyed on the rule or insight subject rather than a row id, and the description is projected fresh each time. That is what lets the queue stay current without losing what you decided about it.
What does Verify actually do?
It depends on the rule, and the method is recorded on the result. It either re-reads the affected pages and runs the rule over fresh markup, re-fetches the single file the site publishes, or reports that only a full analysis can settle this one. Outcomes are passed, partial, failed or unavailable.
Can it push tasks to Jira or Linear?
No. There is no task-tracker integration today, and the product does not claim one.
It Does Not Work Alone
Every part of GrowthHasten reads the same project, which is what lets one of them know something another one found.
Free To Start
Run Action Center On Your Website
Sign up with your domain and the first analysis starts before the redirect finishes. Free plan: one website of up to 900 pages, and three re-analyses. No card, no trial clock.