There is a worse outcome than not being mentioned. It is being mentioned inaccurately — described as the option for small teams when you sell to enterprises, or listed as lacking a capability you shipped eighteen months ago. That paragraph is being read by people who will never contact you to check, and correcting it is not a matter of complaining to anyone.
This is the part of the shift that gets discussed least. Most of the attention goes to whether a vendor appears in assembled answers at all. The more consequential question is what those answers say, because a confident, wrong description circulating during an evaluation removes a vendor from consideration more efficiently than absence does.
Comparison questions dominate the early stage
Someone beginning an evaluation does not search for vendor names. They search for the shape of the problem and then for the shape of the choice — which approaches exist, how they differ, what each costs in effort. Those questions have no canonical source, which is precisely why an assembled paragraph gets built for them.
| Stage | The question asked | Who tends to be quoted |
|---|---|---|
| Understanding | What is this category, what does it solve | Whoever explained it clearly |
| Comparing approaches | Build versus buy, hosted versus on-premise | Whoever named the trade-offs |
| Comparing vendors | How does A differ from B | Whoever wrote it down honestly |
| Qualifying | Does it satisfy this specific requirement | Whoever stated the requirement plainly |
Row three is where most software companies have nothing at all. Writing an honest comparison against a named competitor feels commercially reckless, so nobody does it — which means the paragraph gets assembled from review sites, forum threads and whatever the competitor published. Both vendors then live with a description neither of them wrote.
What a paragraph is built from
Ordering happens at the level of the document in a ranked list, weighing subject match against reputation built up over years. Assembling a paragraph is a different operation altogether: individual assertions are pulled out of individual pages and woven into running text. The thing being picked is therefore a sentence rather than a document — sometimes only the few words surrounding a figure that answers the question.
A checkable claim in one sentence
Named subject, stated fact, condition attached. It survives removal from its page because it never depended on the page.
- Verifiable against something
- Reads correctly in isolation
A claim that leans on its surroundings
"Unlike other solutions on the market, ours handles this more elegantly." Nothing in that sentence can be checked or lifted.
- Unnamed comparison
- Unmeasurable adjective
Nearly everything on a typical vendor site belongs in the right-hand card. This is not sloppiness. Read from top to bottom the prose works fine, and it is exactly what promotional writing has always been. The difficulty is narrower: nothing that lifts a single paragraph and throws away the remainder can make any use of it — and that describes a growing proportion of how software now gets onto a shortlist.
The test is easy to apply and mildly humbling. Take any paragraph from your own site, lift it out of its surroundings, and read it alone. If it still makes sense to somebody who has read nothing else, it can be used. If it needs the sentence before it, it cannot — and the repair is usually replacing a pronoun with the noun it stands for and writing out a condition that was left implicit. Seconds per paragraph, and the highest-return single change available on most vendor pages.
There is a second benefit that is easy to miss. Prose whose sentences stand alone is also markedly faster for a human to skim, and evaluators skim rather than read. The change made for one reason improves the page for the other, which is unusual among interventions of this kind.
What to do about an inaccurate portrayal
Where a wrong description comes from
Almost always from a source that was accurate once.
- A review from three releases ago. Accurate when written, undated, and still the most quotable description of your product anywhere.
- A comparison page written by a competitor. Not dishonest, necessarily — but written from their perspective, and the only comparison in existence.
- Your own positioning from an earlier strategy. An old page describing the product as it was two pivots ago, still live because nobody retired it.
- A forum answer from a frustrated user. Specific, vivid, and about a limitation you removed since.
The only correction available
Publish something more current, more specific and more checkable.
- State the capability with a date. "Available since version 4.1, released in March" is quotable in a way that "we now support this" is not.
- Retire your own outdated positioning first. An old page of yours contradicting a new one is the easiest fix on the list and the one most often skipped.
- Name the limitation you actually have. A page listing only strengths reads as promotional and gets passed over. One that names a genuine boundary reads as reliable.
- Give the page a visible review date. Currency is a signal that undated pages cannot offer, and adding the line costs minutes.
The second bullet in the second block is worth dwelling on. In most companies of any age there are pages describing the product as it was during a previous positioning — still live, still catalogued, still the clearest available statement of something that stopped being true. They contradict the current material, and where two of your own pages disagree, the older and better-established one usually wins.
There is also a limit worth naming, because it defines where effort is well spent. Where a question has an obvious canonical source — a standards body, a platform's own documentation, a large vendor's reference material — that source gets used, and nothing you write displaces it. The productive territory is everything without a canonical source, which in software means comparisons, behaviour under unusual conditions, and anything specific enough that nobody has yet troubled to write it down properly. That boundary belongs in the editorial plan rather than being discovered afterwards.
How to find out what is being said
No report covers this. What exists is asking and writing down what comes back, which takes about a quarter of an hour a month and is done by almost nobody — which is exactly why it produces an advantage.
- Ten questions an evaluator would genuinely ask. Including at least three comparison questions naming your main competitors, because those are the ones that decide shortlists.
- Word for word the same each month. Otherwise last month cannot be compared with this month and the entire exercise produces nothing.
- Record what was said, not just whether you appeared. This is the difference between this exercise and the version most companies run. The description matters more than the mention.
- Note which competitors were named. Over a quarter this produces an inventory of who in your category has published the clearer material.
Keeping those readings beside the conventional figures in one workspace is what stops the habit lapsing after the second month, which is when it normally lapses.
One source type that gets used disproportionately
Encyclopedic sources turn up in assembled paragraphs at a rate wildly out of proportion to how much of the web they represent. Nothing mysterious explains it: the material is broken into labelled sections, it links densely to related entries, and the register it is written in yields self-contained sentences. That makes a placement there worth more per unit than the same placement anywhere else. The cost is not high — ten dollars per placement, available in quantities of none, one, five or ten and nothing between.
| Quantity | Monthly cost | Annual cost | Appropriate where a company |
|---|---|---|---|
| 1 | 10 $ | 120 $ | wants to establish whether it does anything |
| 5 | 50 $ | 600 $ | covers several distinct technical concepts |
| 10 | 100 $ | 1 200 $ | spans a broad platform with many terms |
Nothing sits between the available quantities. Conclude that seven placements would suit you and the decision collapses to five or ten, with six hundred dollars a year separating them. A software firm of this size should generally start with one: for a hundred and twenty dollars it settles the only thing worth settling first — whether this route produces anything at all.
Where a bigger number changes nothing
Behind the second lever sits an entirely different logic. Network placements are priced at a dollar apiece and sold in fixed quantities: none, twenty, a hundred, or five hundred. That quantity records one thing only — the count of placements produced. It carries no information whatsoever about the standing of the sites involved. Buy the top quantity and there will indeed be five hundred of them; what there will not be is a fivefold improvement over the hundred quantity.
Only the encyclopedic lever bears directly on the objective described in this piece. Network quantity moves how heavily a site counts within a conventional list of results; it does nothing to make an assertion easier to lift into a paragraph. Keeping the two entries visible together in the same account at least prevents them collapsing into a single budget line at review time.
None of that is an argument against network placements. They exist for a separate purpose and they serve it. It is a warning about a budgeting habit that recurs constantly: a company resolves to be described accurately during evaluations, then puts the whole allocation behind the option that produces the larger number in a monthly report — which is precisely the option least connected to that aim. Keeping the two justifications apart in a single environment costs nothing and makes the year-end conversation possible.
The comparison page and why it pays
A named comparison, honestly done
Your product against the alternative most often considered alongside it, including the cases where the alternative is the better choice.
- Names the competitor
- Concedes something real
It is the only one that exists
Nobody else writes these, so the paragraph gets built from whatever is available. Publishing one makes yours the available source.
- No canonical alternative
- Reads as reliable, gets used
"We are advertising the competitor"
The evaluator already knows the competitor exists — that is why they searched for the comparison. The choice is whether they read yours or someone else's.
- The comparison happens regardless
- Only the authorship is in question
Concede accurately or lose the page
A comparison that finds your product superior on every dimension reads as promotional and gets discounted entirely.
- Name one genuine advantage theirs has
- Say which buyer should choose it
The fourth card is the hard part and the reason so few of these pages exist. Conceding that a competitor is the better choice for a particular kind of buyer feels like giving something away. In practice it does the opposite: it improves the quality of incoming enquiries, because the buyers for whom the competitor is better stop contacting you, and it makes the page usable as a source, because a one-sided comparison is not a comparison.
If the words never leave the server, none of this applies
A page that assembles itself inside the visitor's browser once loading finishes may well never have its wording treated as wording by anything outside that browser. Marketing sites built on the same framework as the product are not merely at risk of this; it is the standard outcome in the absence of a deliberate decision to the contrary. From inside the company the problem is invisible, since the page looks flawless on every monitor in the office.
A technical review settles whether the text actually ships, in a morning. Whether your own pages contradict each other — old positioning against new — belongs to on-page work, and it is the cheapest correction available. Which comparison questions are being asked in your category is answered by keyword research from measured volume rather than from assumption, and turning those into pages is what content strategy covers. Running all four from a connected account stops them being pursued separately and reaching conclusions that cannot all hold.
One closing note on expectations. Correcting how a product is described does not produce enquiries in the following month. It changes what an evaluator carries into a decision that will be made two or three quarters later, and in a market with year-long cycles the effect surfaces correspondingly late. Judged after a quarter against the number of opportunities created, this work will always look like a failure. The measures that mean something early are different ones: whether the new pages were taken in at all, whether they generate impressions, and whether the monthly readings show the description shifting. The commercial question can be asked seriously only towards the end of the second year — and by then it can also be answered.
Check how your product is currently described
Questions from product marketing and leadership
An assembled answer describes our product wrongly. Can we get it corrected?
Not by contacting anyone. The only available remedy is to publish something more current, more specific and more checkable than whatever source produced the description — and to retire your own outdated pages that support it. That second half is usually where the wrong description is actually coming from, and it is entirely within your control.
Should we really name competitors on our own site?
The comparison is being made whether or not you participate. The only question is whether the paragraph is assembled from your page or from a review site and a competitor's marketing. Publishing an honest comparison, including where the alternative genuinely fits better, is the only way to be the source rather than the subject.
How do we measure this? Nothing reports on it.
You ask the questions yourself and write down what comes back — ten questions monthly, three notes per run, about fifteen minutes by hand. Record the description and not just whether you were named, because the description is the part that decides shortlists. Nothing automated does this job today, and the manual version shows a direction inside three months.
Does being named produce enquiries?
Rarely in any traceable way. When the paragraph settles the question, the search ends there. What it produces is a place on the shortlist assembled during evaluation, which surfaces months later as a direct approach or a search for your company name — neither attributable to the paragraph that caused it.
Should we publish more to improve our chances?
Fewer pages, written better. Twenty thin posts on adjacent topics compete with one another and none of them contains a claim worth lifting. Three pages that genuinely settle a question outperform them and cost less to produce — which runs against every editorial calendar ever drawn up, and is nonetheless the case.
One reference-work slot or a hundred network slots?
For this purpose the reference-work slot: ten dollars a month rather than a hundred, aimed at the source type that actually gets quoted. Judged on ordinary ranking the answer runs the other way. The two do different jobs and each deserves its own justification rather than a head-to-head.