Sunday, 25 July 2010 10:14

Bad data means bad decisions

Written by

A recent poll on Mashable purported to show that users prefer real books to e-books.

Looking at the poll, it's pretty clear: people overwhelmingly prefer real books to e-books.

However, if you take more than quick glance, it reveals its lack of depth. The question had three choices: prefer books, prefer e-books or happy with both.

On the face of it, those seem like valid questions - so what's missing?

There's just one vital question that isn't asked. For this poll to be truly valid, the questions need to be expanded or qualified with at least one pre-question, preferably two:

  • Have you ever used an e-book for a reasonable length of time?
  • Was this e-book like a Kindle (non-backlit, 'electronic paper') or like an iPad (bright, backlit LED screen)?

The problem with the original poll is simple: there's no differentiation between those who have tried an e-book, but prefer real books and those who have never tried an e-book.

The poll suggests that over 40% of people prefer real books, but without knowing what percentage of those users have tried an e-book for any length of time and of what type, the data is disingenuous at best.

It suggests that e-books don't provide a comparable experience to real books - but without the additional questions, that is not necessarily the case.

It's a simple point, but on these simple points, critical business decisions are made. A business might, on the strength of this data, decide to invest more heavily in real books - but in point of fact this poll doesn't tell us any meaningful data about reading preferences.

Relying on bad data or poorly interpreted data leads to critical mistakes.

The first step is always to ask the right questions.

This email address is being protected from spambots. You need JavaScript enabled to view it.