A monitor watches a thing and tells you when it is wrong. That is genuinely useful, it is what most tools in this space do, and it is close to free because the underlying checks are commodity.
A record keeps what was observed, dated, so that a question asked later can be answered. That is a different product, and almost nobody sells it.
The questions do not overlap
A monitor answers: is the certificate valid, is the domain expiring, does mail authentication pass. All present tense, all about the current state, all answerable by anyone at any time from public data.
A record answers: when did we change hosting provider, what was that DNS record before somebody edited it, how long has this been the case, were we running that third-party script during the period of the incident, has any of this been looked after and since when.
Every one of those is historical. None of them can be answered by a monitor, however good, because a monitor's job is complete once it has told you about now.
When each question gets asked
Present-tense questions get asked by whoever is looking after the thing, continuously and casually.
Historical questions get asked at moments that arrive without warning and carry consequences. A handover, when the person who knew has gone. An audit or a security review. An insurance claim. A dispute about what was advertised. A buyer doing diligence. An incident somewhere else, where the question is whether you were exposed on a particular date.
Nobody plans for these. That is exactly why they are hard: the window to have prepared closed long before the question appeared.
Why the record cannot be assembled later
Some of it can, given a few days and a lot of patience. Registrar emails, old invoices, a backup somebody kept.
Much of it cannot at any price. Nobody can tell you what a DNS record contained before it was edited unless a copy was taken at the time. Nobody can produce what a page said on a specific date unless it was captured. And nobody can retroactively establish that something was being watched, which is often the actual question.
This is the property that makes it a different business. A monitor's value is available on demand whenever you decide you want it. A record's value depends entirely on a decision made earlier, and the decision cannot be revisited.
The competitive consequence
Monitoring is copyable in a week. A dozen services already do the checks well and charge very little, and any of them could add whatever feature you thought of tomorrow.
A record with three years behind it is not copyable at all, by anyone, at any budget. Not because the technology is hard — it is a hash chain, which is trivial — but because the input is time, and time does not respond to funding. A better-resourced competitor can match every feature and still cannot hand a new customer a history that was never kept.
That is an unusual position and worth being clear-eyed about. It means the product is weakly differentiated in year one and strongly differentiated in year three, which is the opposite profile from most software. It also means the only strategy that works is starting and continuing, which is unglamorous and largely correct.
Which one you need
If the question you have is whether something is wrong right now, use a monitor. Several free ones do it well and you should.
If the question you might have later is how long something has been true, a monitor will not help, and by the time you know you needed a record it will be too late to have kept one.