In one sentence: almost every tool in this category shows mention rate, citations and competitor comparison, so the feature list tells you nothing — ask instead how each number is produced, because that is where they genuinely differ.
We build one of these, so read this with that in mind. The questions below are the ones we had to answer for ourselves, and several of them we got wrong first.
1. Can I see the question set, and does it contain brand names?
Questions naming your brand will be answered with your brand. That is an echo, not visibility. A set where a third of the questions name you can report a healthy-looking score against a true value of zero — this happened to us, thirteen of fifteen questions, 13% reported against an honest zero. Ask to see the list before you look at any number.
2. Are citations counted from the answer body or the retrieval pool?
Engines assemble a candidate pool much larger than what they cite. Counting the pool inflates citation figures by a multiple — two to four times in our own measurements before we fixed it. If the vendor cannot answer this question precisely, assume the larger number.
3. How are competitors detected, and is it the same code as for me?
Any asymmetry moves in the vendor's customer's favour, which is exactly what makes it sell. Ask whether the same pass, the same matching rules and the same name normalisation apply to every brand including yours, and how ties resolve.
4. What happens to a run that fails?
| Answer | What it means |
|---|---|
| "It shows as a gap" | Correct |
| "It shows as 0%" | Trends will show fake declines |
| "It's excluded from the average" | Averages silently rise; ask if the exclusion is visible |
| "It retries automatically" | Good, but ask what happens when retries also fail |
5. Can I open any number and reach the answer text?
Scores summarise; evidence explains. Without the stored answer, the engine, the date and the cited URLs, you cannot tell a detection bug from a real change — and in this category you will have both, regularly. A tool that cannot show you the underlying answer is asking for trust it has not earned.
6. What does it cost to rerun, and how often can I afford to?
Quota is the binding constraint. A tool with a beautiful interface you can only afford to run monthly gives you twelve data points a year against a metric that varies between runs. Work out the real cadence at your budget before comparing features, because it determines whether you have a measurement or a series of anecdotes.
The honest summary: if a vendor answers all six precisely, the product is probably sound whatever the interface looks like. If any answer is vague, that is where the number is being made.