A single-product review peaks in its first fortnight and declines steeply afterwards. A comparison covering the same products often grows for a year or more. The difference is in the question asked.
The two videos catch different moments
An individual review answers whether a product is any good, which is the question asked when it is newly announced and largely unknown.
A comparison answers which of two known options to choose, which is the question asked immediately before a purchase, at whatever point that purchase happens to occur.
Purchases are spread across the whole life of a product, while curiosity about a new release is concentrated at launch, so the two videos are fed by very differently shaped demand.
Search behaviour reinforces the gap
Queries naming a single product tend to arrive during the announcement window and fall away as the news cycle moves on to whatever replaced it.
Queries naming two products together are made by people who have already narrowed their options, and that narrowing continues long after launch coverage has ended.
The comparison also matches more query forms, because any of several phrasings pairing the two names can lead to the same video.
Comparisons survive partial obsolescence
When one product in a pairing is replaced, a single review of it becomes largely historical. The comparison retains value because the other product is often still current.
Discontinued models also move to a second-hand market where buying decisions continue for years, and those buyers are asking exactly the comparative question the video answers.
This gives the format an unusually long useful life relative to the effort, since it was produced from testing already done for the individual reviews.
The format demands a stated basis
A comparison only works if the criteria are made explicit. Without them the video becomes a preference statement, which ages badly and satisfies nobody who weighted things differently.
Setting out the tests, then reporting each product against them in the same order, allows a viewer to reach a different conclusion from the same evidence.
That structure is also what makes the video usable by someone who cares about one criterion only, which broadens the audience considerably.
Where the format fails
Comparisons between products in different classes rarely find an audience, because nobody is choosing between them and the query is therefore never made.
Pairings involving obscure products fail for the same reason. The durability of the format comes from the volume of people facing exactly that decision, not from the format itself.