What the customer needed..

It's pretty much inconceivable if you work in software that you haven't seen this picture posted up in your lunch room at some point. It is of course a witty look at why the software process doesn't always produce the expected functionality in the most elegant and economical fashion. Depicted are a variety of parties that must share the blame for how the end product differs from what the customer needs.

For the most part it's a reflection on communication and how personal perceptions lead to differing end results. However, when I look at it, the most striking point, and the one that I feel is most overlooked, is that the customer's description and what they needed (first and last panels) are subtly, but distinctly different. This I feel is the problem with lots of software design - the customer does have a concept, a feeling, for what the software should do, but finding a customer who can vocalize this into specific functionality rarely, if ever, happens. Turning to the customer to provide feature lists or functionality is an easier way out than stepping back and making a decision about the business problem that is being solved. What does the customer really need? What do the potential customers I have yet to sell to really need? How can I quickly, economically build that, and not get sidetracked into other features that are affiliated but take away from my concentration on the main business problem?

Is it just me? (software defaults)

I don't know how the wrong decisions about software defaults are made. Take the example on the left that happens to be from Excel 2007, but persists since the dawn of Excel. The bar chart is the default drawn when you choose the data shown.

To me it seems nonsensical that the categories top to bottom are in the opposite order to that of the data, and that the order of the series is reversed (the yes bar should be on top).

I can reverse the order by choosing reverse category order in the axis format menu, but that also switches the horizontal axis to be at the top - you need to format it again to move the axis back down. Perhaps they did a user-group study and I'm the odd one out - the expected behavior is that the order is reversed. If not though, was this a decision that was made (wrongly) by a single person or software team? Is it a bug that no-one ever noticed? Was it understood to be wrong in an earlier version but not corrected due to concerns about compatibility? Am I just too fussy to think that defaults should reflect what the majority would expect the default to be?

Old news

A few days ago I talked about the gradient effect on the Excel in-cell bar charts. Coincidently today I came across a post I made to the Dashboard Spy blog back in the heady days of May 2006 when Office 2007 was in beta release. I posted a quick review of some of the conditional formatting available, including criticizing the gradient effect, but liking the ability to shrink the charts to make effective sparklines.

I was clearly too blown away by Office 2007 (or too confused by the ribbon?) to notice that even in my example shown, the bar chart scale does not start at zero, giving a false impression of the variability of the data. Anyway, maybe I was one of the first to vocalize my dislike of the gradient...

On a related note, a commenter on another website in Sept 2006 gives this possible reason for Microsoft's decisions on the in-cell bar charts:

The default graphs were hideous at best, but now, thanks to a focus-group-tested and user-centric decision probably made by marketing drones without a brain: Microsoft Excel deliberately misrepresents data, because it turns out, users didn't like empty cells in bar-graphs; idiots :-)

Undeniably harsh and unfair, but the commentator has some points that I will touch on soon about user-groups and designing interactions.