How we rate plans

The calculated value score

The calculated value score is for owner-led small businesses. It compares what each plan costs a month to do 10 questions a week on ChatGPT and Google AI Overviews, for 1 brand, 1 country and 1 person signing in, paid monthly. The calculator shows it as a small starting set of 10 questions.

• We cost each plan by the calculator’s own rules, with any extras it needs.
• What the workload needs is a requirement, which a plan meets or does not. More people, brands, AI services, questions or checks than it needs earn nothing.
• A limit on countries, brands or people signing in that the tool maker does not publish counts as 1, as in the calculator.
• A plan paid monthly with a 12-month commitment counts as paid monthly, and its score says so.
• A plan that also buys other tools, such as SEO tools, is rated on its full price.

The cheapest plan that does the workload scores 100. Every other plan scores 100 times the cheapest cost divided by its own, so a plan that costs twice as much scores 50. Scores are rounded to a whole number before they become stars.

Only plans priced in US dollars are compared, and the cheapest paid one sets the comparison. A plan that does the workload for nothing scores 100.

How a score becomes stars

Score	Stars	Cost against the cheapest plan, roughly
80 to 100	5 out of 5	Up to 1.25 times
60 to 79	4 out of 5	Up to about 1.7 times
40 to 59	3 out of 5	Up to 2.5 times
20 to 39	2 out of 5	Up to 5 times
Below 20	1 out of 5	More than 5 times

A plan is not rated when it:

• cannot do the workload, or might not
• is priced on request, or sold only yearly
• is priced only in part, such as a starting price
• is listed for agencies
• is priced in another currency, which we do not convert

The value for money page gives each reason.

Help acting on the results

Some tool makers say their plans give recommended actions or tasks, or help drafting the work. The comparator shows their words, marked untested. They do not count in the calculated value score.

A product’s tasks and drafts are scored only once it has been through the practical test below, run the same way for each product. No product has been through it yet.

The practical test of tasks and drafts

The scenario: one owner-led small business with one website, on the score’s workload, with the same 10 questions for every product. We publish the questions with the result.

1. Use the cheapest plan that does the workload, at its published price, with no help from the tool maker.
2. Enter the website, the brand and the 10 questions.
3. After the first full week, record the plan’s tasks or recommended actions, up to 10.
4. Ask the plan for a draft for each of the first 3 tasks, where it offers drafts.
5. Save a dated screenshot of each task and draft.

Each task scores 1 point for each yes:

• Does it name the page, site or change to make?
• Does it point to the AI answers or sources behind it?
• Could the owner do it in a working day, without specialist help?

Each draft scores 1 point for each yes:

• Is it free of wrong facts about the business?
• Could it be used with light edits?
• Does it do what its task asked?

The result is the share of points scored, for tasks and drafts separately. It is shown beside the calculated value score and does not change it, so a product we have not tested is not marked down.

Editor’s ratings

An editor’s rating is our judgment of how useful a product is overall for a named kind of buyer, from our own use of it, recorded with its plans, tasks, dates and limits. It sits beside the calculated value score, never in its place, and our formula does not produce it. Discoverable is the only product with one. Its page says what we used, what we found and the limits.

Changes to how we rate plans

Changes to how we rate plans

Date	Change
October 29, 2026	The score now prices 10 questions a week, the workload we chose for owner-led small businesses, in place of our 25-question Small business example. We made this change after seeing the scores of October 28.
October 29, 2026	Where a tool maker states a count two ways, a workload within what both readings allow now fits. This rates Discoverable Scale and SE Ranking Core and Growth.
October 28, 2026	The score compares what each plan costs to do the Small business example.
October 28, 2026	People signing in, brands and AI services became requirements of that workload, not points.
October 28, 2026	Help acting on the results is shown untested and no longer counts.
October 28, 2026	Editor’s ratings were added.
October 27, 2026	Help acting on the results was added at 10% of the score. Price per 1,000 answers went from 60% to 50%.