Your data audit ends up here

A blog by Martin Erasmuson.


 “We’re going to do,” said the manager “an organisation-wide data audit.”

“If you like, yes,” said the Information Architect.

“Will that help?” asked the manager.

“No,” said the Information Architect, and gave him a friendly smile.”

― Martin Erasmuson, Information Architect (adapted from ― Douglas Adams, The Hitchhiker's Guide to the Galaxy)


 The org-wider data audit.  It seems to be the right thing to do.  Get a handle on where all your data is.  And let’s not beat about the bush, there’s a lot of it.  OK, that’s still beating about the bush, your organisation has data for Africa.

 So, the data audit starts. It takes ages, months, and a few thousand hours.  It turns out there is data everywhere, in every conceivable format; even some inconceivable ones.  And the big day arrives, with much fanfare, the results:


 But what does it mean?  In a salient moment, the manager recalls the conversation, many months previously with the Information Architect: “Will that help?” and the answer; “no.”

 What is wrong with this list?  It has no context.  We can illustrate that with a simple question: What is the most important item?  Without context, we cannot answer that question.  Or we decide importance based on our own cognitive bias.  What does your brain tell you about importance if we add some content?

The goal is to make a fruit salad!

We want to post many fragile items!

 Without context, the list of eight items above; and your organisation-wide data audit, are meaningless.  Imagine running a restaurant like that where your customer strategy is to keep turning up at the customers table with random dishes until eventually, you luck onto one they want.

 The ingredients you stock in your pantry is determined by the dishes you want to serve.  If you don’t need it to produce a dish that a customer wants, you don’t need to stock it in the pantry?

 Sadly, most organisations are doing just that with their data i.e. collecting and maintaining ingredients (data); with all the overheads that go with it, that do not contribute to any dish i.e. an information product supporting an organisational outcome.

 The only thing that adds context to your organisations data strategy; what data you actually need, is to first work out the information outcomes you want, and THEN the data you need to produce them. 

That exercise typically reveals three truths about your data management:

1) The data you need and are collecting and managing

2) The data you don’t need but are collecting and managing for no reason

3) The data you need, but currently do not have at all


This sort of makes you data strategy pretty simple: Keep doing 1; stop doing 2; and start doing 3 (either capture or purchase the data you need).

 Infonomics practices start to add layers of subtlety to that, but the overall concept remains unchanged.