It doesn't matter whether you are designing a new piece of software, a business process or even your end product, the importance of applying common sense and considering real-life scenarios simply cannot be understated.

A mistake can be very costly both in terms of time and reputation. If that mistake affects the way your customers interact with you, it could both expose your business to ridicule and potentially lose you custom.

Take the recent case of Southern Electric quoting a customer £13,088.43 after they entered a meter reading lower than the estimate on their bill. The companies spokesman explained that the issue occured due to a 'feature' in their website.

They only allow a lower reading to be entered for two weeks after the bill is generated, after that the system assumes it is a new reading and generates the bill accordingly (in this case by taking the meter around the clock).

So in this case, the user believed he was going to save around £20 by entering the reading but instead was given a much, much higher amount: hardly the reliable service yor customers want!

Planning and Common Sense

Quite why the arbitrary limit of two weeks was even imposed is unclear, but even if we give the company the benefit of doubt the mistake is still pretty large;

On an Imperial meter such as the one in use (4 digits on the display), 1 unit is equivalent to approx 2.832m³ of gas used. The estimated reading was 7328 and the customer entered 7305, so the bill generated assumed 9976 (28,252m³) units had been used.

This type of meter has a maximum flow rate of 6m³ per hour (anything higher would indicate a leak), so that level of usage would require 4,708 hours (196 days) of maximum flow!

The spokesperson said that a bill for this amount would never have been generated, but it doesn't explain why their systems weren't designed to automatically detect such anomalies. The customer couldn't possibly have used that amount of gas, and a sane designer would compare the usage and assess the probability that the reading was in fact a lower adjustment.

Without even needing to know the maximum flow rate of a meter, a basic form of logic would have been to assess the proximity of both the estimated reading (7328) and the inputted reading (7305) to the rollover point (9999) and then decide whether or not it was likely the meter had been 'around the clock'.

The Potential Fallout

At first glance, the potential consequences of this case seem quite low. After all, the company has stated that the bill would have been flagged for attention before it was sent out, so where's the harm?

The main issues that could arise are two-fold;

• It's degraded customer faith in their systems (what else could they be miscalculating?)
• It's plausible that it could have been missed.

The latter is particularly concerning, given the increasing reliance on Direct Debits. Whilst a customer receiving a £13,000 bill is almost certainly going to phone you and query the amount, if that money has already been deducted from their account a whole new issue is presented. Refunds take time, and the customer is not going to be feeling forgiving when the issue is your fault!

Identifying Troublesome Areas

It's very important that you do identify areas that have the potential to cause issues, and act to resolve them. Whilst work is underway, you need to make sure you are telling customers about the issue so that it can be more easily resolved.

In this case Southern Electric stated that it was a 'feature' and that their site will only accept a lower reading for two weeks after a bill. Unfortunately, their site states

"This is our flexible meter reading and bill service. Here you can enter your meter reading service whenever you wish"

Without actually mentioning the two week limit, so it's unsurprising that customers have been caught out in this way. As the comments section of the linked article show, it's not the only mistake energy companies often make, which then leads to the quality of their customer service departments being brought to the forefront (would you be happy to spend 30 minutes on hold as a result of someone else's mistake?)

Real-life application

It's unclear why their site was designed in this way, it could be the result of cost-cutting, an incompetent developer or even because that was the way they wanted it to work. Where they truly failed was in considering the real life uses of the system, it should have been obvious that people will sometimes enter readings lower than their estimate, and placing a warning stating that there was a two week period in which to do so would have been an obvious move (or even, not imposing the limit at all!)

Learn their lessons

Every business has to plan, whether its an event or a product. Never rely on assumptions without seriously considering how the target audience are likely to behave, especially when designing anything with a financial edge to it. Don't allow your company to become the one who accidentally billed a customer £1000's.

Consider the impact of every change within the project management lifecycle and never forget that users generally don't use products the way in which they have been designed to work. Whether developing a product yourself, or outsourcing the work, consider every possible angle and avoid imposing arbitrary limits rather than employing a little bit of logic.

And the final, and potentially most important, lesson is to ensure that you have people available to help quickly resolve any issues that do arise. It's bad enough to make a mistake that affects a customer, but to then make them wait whilst you find someone to deal with their issue is something that customer will never forget!