Date: 11th February 2009
Having worked on/in the Internet now for 16 years, sometimes it feels like everything should ‘just work’ by now.
Back in 1993 I ran an ISP where we had only 256K bandwidth to the world, for our entire UK customer base to share! Now, we all get 2M, 8M and more to the home: and our ISPs hookup in Gigabits to each other.
Hardware is cheaper. The web server software out there is now more mature.
Why can’t it all just work, all the time, like it should?
But of course, that 16 years has been characterised by huge technology shifts – so actually, the web sites we all use are actually ‘brand new’ and pretty much seat of the pants.
This week for example, we’ve been helping a retailer who reckons that the ROI on our work planned for the whole next 12 months, has been recovered in the first month!
The reason of course, is that it’s not hard to find significant sources of lost sales on the websites of even the bigger and better players. Because things, and especially technology, don’t always ‘just work’.
We set out to do what we always do – helped the client to map out the money-making, vital User Journeys that their sites offer: the multi-page routes by which customers find and select products, Checkout those products, log back in and check the status of an order and so on.
We had those on the table, and then we got input from the call centre as to what journeys they felt users were most prone to complain about. And we asked the tech team, if there were any particular blocks of technology behind the scenes, that they felt were perhaps the least solid and reliable; what things had been problems in the past few months.
All that resulted in us constructing 7 or 8 different multi-page User Journeys
And then we started measuring the Journeys 24/7 -using our tools to ‘mystery shop’ the client every 5 minutes, and following the journeys each 5 minutes, and reporting when something doesn’t work right: an error page, wrong content, a product apparently out of stock when it isn’t, when the basket is empty after you’ve already put something in it, and etc
And within the first month, we’ve identified a problem that’s been quickly fixed, and will save the equivalent of 3 hours prime time shopping a month .
It wasn’t a problem of sites crashing, or horrible error pages – just a subtle software bug that meant for a small percent of users Checking out, that their basket would from one page to the next become empty. It was to all intents a valid page, but saying ‘nothing in the basket’ – it wasn’t the right page for that point in the Journey. And real users got thrown off course, and gave up on the CheckOut.
There were no errors in the tech team logs, because the page was OK; and likewise the web analytics too had no hint.
Now that it’s fixed, that’s going to mean that the ROI for all their future marketing campaigns is going to be better too – as less visitors will drop off along the way: so the Marketing team Success are going to look like stars for a few months!
And looking back at some recent campaigns, they now realise that some with a low ROI, where actually suffering because of this particular root cause, so they need to revisit some of those previous tactics, to see if they’ll now work as expected.
It’s funny, back at the start, the client’s Marketing folk were reluctant to get into measuring of User Journeys 24/7; somehow this felt like it should be the tech teams area, those guys said they had 24/7 server monitoring running after all, shouldn’t they know about sporadic errors coming from software or technical root causes?
Of course real users don’t hit ‘servers’ they follow multi-page journeys – and if that’s not being measured, then it’s not being managed, and if it’s not being managed, then there’s opportunities to reclaim some of the lost sales.