You see what you expect to see…

June 26, 2017

I have a confession to make – one that is a bit embarrassing. As a researcher, I know you must not trust your own opinions but should look at the evidence, because often our instinctive view of world – or of people’s behaviour – is wrong!

A while ago I was caught out ignoring the evidence of my own eyes. I had travelled north from Greece to Croatia and, because they are on the same longitude, had assumed they would be in the same time zone. When I saw a Croatian church showing the time was an hour earlier than I thought I assumed the church clock was wrong. It took me 24 hours and several clocks to realise I was at fault not the clocks!
The problem was my expectations. I had a firmly held view on ‘the state of the world’ and I chose to ignore the evidence that contradicted this. This is a well-known cognitive bias.

I saw another nice example of this bias when watching some testing recently. The client sells a wide range of products both through stores and online. The offer can be quite complex. Sometimes the items may only be available online – ‘Web specials’ – sometimes only in store if, say, they are too big to easily transport.

They were trying to work out how best to ‘message’ this complexity, so we tested a range of different versions with users. These used different words and different visual presentations.
What was interesting is that often the messaging had little or no effect on users’ understanding of the proposition. They saw what they expected to see. Users who shop online arrived at the site saw lots of items they could buy online and all the iconography of an ecomms site (e.g. baskets etc.) – so assumed all products could be bought online. They managed to ignore, or more likely not process, large conspicuous ‘In store only’ messaging. Similarly, users who expected that all products would be available instore would happily ignore ‘Online only’ messages. Observation of the session videos show the eye tracks looking at the message, but when asked if they could buy online or instore their responses often contradicted the messaging.

What to do? The conclusion was that the messaging was not helpful and not to worry about it – certainly it was not sensible simply to make the messaging ‘shout’ louder. Better to pick up the issue with error messaging at the ‘add to basket’ stage.

The broader conclusion is of the need to work with users’ expectations of how something will work – because you will have to work really hard to change this. A bit of embarrassment means this is a lesson I have learnt well!