In the original 2020 article, a client planning session produced 28 possible tasks before the team had even mystery-shopped the site. Finding more work was easy. Choosing what to do first was the useful part.

Name the outcome before choosing the task

Be specific about what the business needs: more qualified enquiries, fewer abandoned purchases or a working mobile buying journey. Then ask what evidence connects each proposed change to that outcome.

An untidy configuration file and a broken purchase button are not equivalent simply because both appear in an audit. A neat checklist is not the commercial result.

Compare impact, effort and confidence

For each candidate, record:

  • The customer problem and the evidence that it exists.
  • The outcome the change should improve.
  • The work required, including testing and dependencies.
  • How confidently you expect it to help.
  • What you will check after release.

A short, well-supported fix can deserve priority over a large speculative project. But quick is not automatically valuable. Removing a plugin because it appears slow still needs a check that customers do not depend on what it does.

Keep essential safeguards outside the sales queue

Security, privacy, accessibility and serious reliability problems may need action even when you cannot attach immediate revenue to them. The original article’s revenue-first argument is not a reason to defer those responsibilities.

An empty robots.txt file is not automatically a fault, but when evidence shows it is blocking legitimate crawlers (for example, after an audit flags zero indexed pages), it becomes a prioritised fix because the outcome is lost visibility rather than a cosmetic preference.

Check whether the change helped

Release a change you can identify, verify the customer journey and compare the relevant results. Use device and channel detail where it matters. Do not attribute every movement in revenue to the latest release.

See why store-wide conversion rates can mislead, indexing diagnosis and our conversion optimisation service.