Two extensions in my category ship under the same name, and one product ships under two. I stopped matching on names

In July I downloaded eight browser extensions that compete with mine, opened each package, and read
the manifest file the developer ships inside it. That file declares what the extension may do and
which domains it may read and modify. I published the table.

Since then I have been writing to authors of comparison articles, and every letter contains one
sentence of the form: this many of the tools you cover are in my table. It sounds like the easiest
sentence in the letter. It is the one that keeps being wrong.

Three collisions, in one category, in one month

Two different extensions ship under the same product name. One of them is in my table, anchored by
the domain its manifest declares. The other is a different company's, linked by a review site to an
entirely different domain. If you match on the name, you claim to have measured something you never
opened.

A second product ships under two names, having been rebranded. A comparison article listed it under
the new name. My table listed it under the old one. Counting names, it looked absent. It was there.

A third product has a name close enough to both that I spent a cycle deciding it was one of them. It
is not related to either.

What I was doing wrong

I was asking whether a name in someone else's article matched a name in my table. That question has
no reliable answer, because names are marketing artefacts. They get rebranded, they get reused by
unrelated vendors, and nobody is obliged to make them unique.

What I have that is not a marketing artefact is the manifest. An extension must declare the domains
it touches, and those domains are its vendor's. So the question became: which row of my table
declares the domain this article's link points at?

That question has one answer, or none, and either is useful.

The rebranded product settled itself the moment I read my own row without truncating it: the manifest
declares both domains, the old one and the new one, because the extension still talks to both. My
own measurement had held the answer for a month while I went and asked the vendor's website.

Two counting traps that came with it

Counting occurrences of a product name in an article is worse than useless here, and I have the two
failures to prove it.

One product name is a substring of another. Searching an article for the shorter name returned six
hits, and the article does not mention that product at all. Every hit was inside the longer name.

On another page the same search returned one hit, which was inside a hashtag naming the topic rather
than any tool. One occurrence, zero mentions.

Both times the number was real and the conclusion drawn from it was not.

What I do now, in four lines

Read the article's link for each tool, not its label. Take the domain. Ask which row of my table
declares that domain. If no row does, say so, and say that this proves the name does not match rather
than that the tool is absent. Then write the sentence with the number that survives.

That last part matters more than it sounds. My letters now say things like one of the six, where they
used to say two of the three, and the smaller honest number is easier to defend than the larger one I
could not.

Disclosure

I build BlueTicks for Gmail, a Chrome and Firefox extension that shows WhatsApp style ticks in your
Gmail sent list, one tick sent and two blue ticks opened. It costs 4 dollars a year, and the free
tier covers 30 emails a month. Everything above comes from distributing it in public and writing
down what the tools did. You can find it at blueticks.io.

I have not named any of the products here, and that is not delicacy. The names are the part of this
story you cannot rely on.