What an automated blog cannot do, from the people selling one

Every vendor in this category publishes a post about why their product works. Far fewer publish the list of things it does not do, which is unfortunate, because that list is what a serious buyer is actually trying to assemble.

So here is ours. These are not hedges. They are the reasons we have told people not to buy from us.

It cannot know anything that is not written down

The engine reads your public material and whatever you give it. It has no access to the conversation your founder had with a customer last Tuesday, the reason you killed a feature, or the thing everyone in your company knows and nobody has ever typed.

That undocumented knowledge is usually the most valuable content you have. Getting it out requires a person sitting with another person and asking questions, which is a human activity we can prompt but cannot perform. Companies who treat onboarding as a form to fill in get a blog that reads like their website. Companies who treat it as an interview get something better.

It cannot have an opinion you have not authorised

It can argue a position well. It cannot decide which positions your company is willing to defend in public, and it has no way to know that the sentence it just wrote contradicts what your sales team promised a large customer in March.

This is precisely why we insist on approval and why we quote ten minutes a week rather than zero. A vendor offering a fully hands-off blog is offering to publish unauthorised opinions under your name, which is a strange thing to want.

It cannot do original reporting

It cannot interview your customers, run a survey, sit in on a support call, or go and find out something nobody has published. Content built on original data is the most durable kind, and it is the kind we cannot manufacture.

What we can do is turn data you collected into a good post. If you run a survey, we will write it up well. We cannot run the survey.

It cannot fix positioning

If a company cannot explain what it sells and to whom, its blog will express that confusion at a higher rate. Automation is an amplifier, and amplifying an unclear message produces a great deal of unclear message.

We have twice recommended that a prospect spend their first three months on positioning with someone who does that work, and come back afterwards. Both times it was obviously the right advice and obviously bad for the quarter.

It cannot rescue a product nobody wants

A blog changes how many of the right people find you. It does not change what happens when they arrive. If the underlying offer is weak, better content produces a more efficient demonstration of that weakness.

This sounds obvious written down. It is not obvious in a meeting where content is being discussed as a growth lever, and it is worth saying out loud before money changes hands.

When you should not buy this at all

If you publish twice a year and that is genuinely enough, do not automate it. Write the two posts.

If your content is regulated to the point where every sentence needs legal review, the approval step swamps the saving and you should hire a specialist who understands the regime.

If you have a good writer in-house who wants to write and has the time, keep them. We are not better than that arrangement; we are better than the arrangement most companies actually have, which is a good writer who has no time and a calendar nobody owns.

And if what you want is forty posts a month because a competitor publishes forty posts a month, we will decline. That is the failure mode Google's policies are aimed at, and it does not stop being a bad idea because it is cheap.

Why publish this

Partly because a list of limits is the most useful thing a vendor can give a buyer, and it is rare enough to be worth something on its own.

Mostly because every item above is something a client would discover in month three anyway. Finding out then, having paid, is how a supplier relationship goes bad. Finding out now, in public, is how it starts well.