Measure Success

How do you measure success?

Can a bad process be a good process?

How bad are targeted ads on the web..?! I read a great article about how feminism is more important now than ever so I shared it on Facebook. A few hours later and I’m getting lots of web ads trying to get me to join woman’s business groups.

Another time I shared an excellent article written by a teenage Muslim talking about their experience as a minority group in Australia and yep, a few hours later, I’m getting ads for ‘Converting to Islam’ and ‘Start Teen Modelling’.

I’m neither a woman, a teenager, someone looking to convert to Islam or someone hoping to become a teen model.

Go home targeted ads – you’re drunk.

Clearly, the accuracy of these targeted ads is appalling.I’m struggling to think of a time where I might have clicked on one.

Whilst I was thinking about just how poor these targeted Ads perform – it got me thinking. Maybe they are hugely successful despite being so bad?

It all comes down to what constitutes success.

I remember back in the late ’90s when early voice recognition tech started hitting call centres. All the directory assistance services numbers (think Yellow Pages or White Pages) started to implement robotic voices asking you to “Please state the name of the company you are looking for”.

You’d say “Louisa’s Florist” and the computerized voice would answer…

“Did you say ‘Jims Mowing?”

Uhh… Nope.

The tech was terrible. It was a horrific customer experience. You got used to giving it a go first (just in case) and then pressing whatever to get to an operator.

I remember thinking back then just how bad this tech was.

But I read an article talking about what a raging success the technology was. Basically, before the internet (can you even remember what that was like?), if you wanted to look up the phone number for anyone or any company, you’d “let your fingers do the walking” and look in a phone book.

Remember these…?

Or you’d call directory assistance. These call centres used to get hundreds of thousands of calls per month, or even per week. They’d be quick calls, but the volume was staggering.

So when the article said that the new voice recognition tech had a 5% success rate, this translated to thousands of calls being answered every day without human interaction – and therefore, with no cost.

So, depending on how you measure success – it was hugely successful.

I think targeted web ads must be going through something similar. It’s appallingly bad – but perhaps it’s better than blanket ads that are not targeted at all, so from that perspective, it’s very successful.

But both examples from the user experience – not so great.

Be careful in your Lean improvement projects that when you are measuring success, that you’re taking a holistic view and not just looking at the result you were hoping to improve.