Feedback tool feature matrix, thirty two vendors on nine questions

This is the capability half of the category as a file. Thirty two vendors, nine questions each, read on the vendors' own pages on 14 September 2026.

Every cell says one of three things, and the third one is the reason this file is worth more than a tick chart. Either the capability is there, with the plan tier that gates it written into the cell. Or the vendor's page says plainly that it is not. Or nobody could read an answer, which is recorded as unpublished.

Unpublished is not a no. An absence of evidence on a marketing site is never proof that software cannot do a thing, and a matrix that quietly turns silence into a cross is the thing this file exists to avoid.

What is in the file?

Sixteen columns and thirty two rows.

The first names the vendor. The next three are commercial: the entry price, the unit it is billed on, and what the free tier holds. The next nine are the capabilities below. The last three are the date the row was read and the two addresses it was read at, one pricing page and one product page, so any row can be checked rather than believed.

Take the file. It may be copied, quoted and used commercially with no attribution and no link back.

None of it needs attribution. If you would rather point at it than copy it, the address is https://votefirst.app/en/data/feedback-tool-feature-matrix and the figures on it were read on 14 September 2026.

What do the nine columns count?

Read the second column and the third together rather than either alone. A capability that twenty one vendors publish and eleven say nothing about is probably near universal in the category and simply not worth a line on eleven marketing sites. A capability that two vendors publish and six explicitly refuse is a real fault line.

Which column actually separates these products?

The last two, and they separate them in opposite directions.

Self hosting has six vendors stating plainly that it is not available, which is unusual: this category prefers silence to a no. That makes it the one question in the table where the answers are mostly known, and the answer is mostly no.

Revenue weighting is the other one, and it is the column where reading the cell matters more than reading the count. Five vendors say in their own words that a vote can be weighted by what the customer pays. Four more do something close enough to be listed and different enough to matter: they sort, filter or prioritise a list by a revenue figure without changing the vote itself. Those two things get described with the same words on a comparison site and they produce different answers when two requests have twelve votes each.

Two vendors publish a no, one of them marking the feature as coming soon in its own pricing table, and twenty one publish nothing at all.

Why is there a dagger in some cells?

Thirty four cells carry one, and it means the answer came off the product page rather than the pricing page.

That distinction is worth keeping because the two pages disagree more often than you would expect. A pricing page is a commercial document and tends to name only what gates a tier. A product page is a marketing document and tends to name everything, including things that turn out to be on the top plan. Where a cell has a dagger, the tier it names is the least certain part of it.

How was it built?

By reading the pages, one vendor at a time, in a browser.

Nothing here came from another comparison site, a vendor's own comparison page, or a directory listing. Those three sources are where wrong figures in this category propagate from, because each copies the last and none carries a date.

The script that renders the file is vendor-matrix.py in this project's own research directory, and it renders from a ledger that is re-read rather than edited, so a vendor who changes a plan changes this file on the next sweep rather than accumulating a correction on top of an old row.

What this file is not

It is not a scorecard and the columns are not weighted against each other.

Nine capabilities is not the whole of what separates these products, and a vendor that publishes eight of nine is not therefore better than one that publishes five. Most of what decides this purchase is the unit on the invoice, which is the other file, and what the board feels like to the customer who has to use it, which no table can hold.

If you want these nine read against the things a security review asks for, the enterprise shortlist does that. If you want the prices rather than the capabilities, forty one pricing pages are grouped by what each one does.