Two posts, measured

Blog

Writing about how Consonas is built and what it costs to build it, including the parts we got wrong. Not content marketing.

There are two posts. That is not an accident of a new site so much as a deliberate rate: we write when a decision has been made and measured, rather than on a schedule, and decisions worth writing up do not arrive weekly.

Both are the same choice approached from opposite ends. The first explains why every organisation using Consonas gets a database of its own instead of a share of one. The second is what happened when we measured one of those costs instead of estimating it, and found a mechanism we had built for a problem fifty times smaller than we believed. It is the one we would rather you read, because it makes us look worse.

Posts

Everything written so far

How we write

What this blog is held to

Checkable against the two posts that exist instead of a statement of intent.

Where a post makes a quantitative claim, the measurement is in it along with enough of the method to repeat it, because a number without a method is an adjective wearing a costume. Every architectural decision has a bill attached, and the bill is the interesting half.

Every post is written by somebody who made the decision it describes, which rules out guest posts and sponsored ones by construction, and there is no comment section, because an unmoderated one on a supplier's own site becomes a place where the supplier appears to endorse whatever is left standing. Nothing here exists to rank for a phrase, which rules out listicles, comparisons with competitors we have not used properly, and case studies with the difficult parts removed. That is a commercial cost and we accept it.

Nothing here announces a feature that is not built either. The roadmap is the honest place for intentions, including what has been decided against. Likely next, with no date attached: measuring a marketing site rather than looking at it, against the Web Content Accessibility Guidelines and the rest of what the accessibility page covers, and why the test environment is the most dangerous thing you own.

Choosing

Which of the two to read, and who should read neither

Read the database post first and the free plan post second, because the second is the bill for the first. If you have to answer where the data sits and under whose law, the short version is that there is no United Kingdom region: an organisation is created in the European Union or the United States and stays there. The sub processor list and the rest of the legal pages carry the version we are actually held to.

Neither post tells you what the product does, or what to type on Monday. The product pages describe what is built, the pricing page says what each plan includes and where the limits sit, the help centre starts with getting started, and the customers page is the one to read if what you are really weighing is whether to trust a supplier this small.

Corrections

What happens when a post stops being true

A post keeps its original figure where it stands and gains a dated note at the top carrying the new one, the date it was taken and what moved underneath it. The old figure is never edited in place, because a reader who quoted the post last year would then be holding a quotation of something that no longer exists. Where a reader found the error, the note says a reader found it, and where a correction makes us look better it gets the same note as any other, because a policy applied only when it hurts is a marketing decision wearing the costume of a principle.

Everything in both posts is self reported, and you cannot open the repository to check any of it. Read the method rather than the adjectives, and treat a supplier's account of its own carefulness as a claim rather than as a fact.

Blog questions

Asked about the writing rather than the product

Is there a feed or an email list?

Neither. A feed for two posts is a mechanism with nothing to carry yet, and an email list would mean holding your address for years so that we can send you something twice. If we add one of the two it will be the feed, because a feed holds no personal data about anybody, and a bookmark costs you nothing in the meantime.

Are the numbers in the posts audited?

No. They were measured by us, on our own system, on a synthetic organisation, and no third party has checked any of them since. Every engineering blog you have read is in the same position, and most of them do not say so, which is the reason it is worth saying.

Does anything written here commit you to anything?

No. A post explains a decision as it stood when the post was written. What we are held to is the terms, the pricing page and the roadmap, and where a post disagrees with one of those, the other three are the ones that count. If a post ever reads like a promise, treat that as a defect in the writing rather than as a commitment you can rely on.

Or just use the thing the posts are about

A free organisation, with its own database, in the region you choose, in about a minute.

Three people, a thousand relationships, no card and no time limit.