Comments on: Test and Target silly season has arrived

By: Tim Elleston

Tim Elleston — Thu, 29 Mar 2012 23:05:17 +0000

You’re right, it was a bit of a nightmare scenario, given the requirements, but also a triumph in terms of the platform being able to handle it and produce actionable results at each of the sub-campaign A/B test levels. They had definitive winners across all of the campaigns. So while complex, actionable results came through.

By: Anonymous

Anonymous — Wed, 28 Mar 2012 21:07:18 +0000

How is any of this actionable? This seems like a nightmare scenario, not a triumph.

By: Elusive engagement – Part II – Visitor scoring | Elephants and Analytics

Elusive engagement – Part II – Visitor scoring | Elephants and Analytics — Mon, 20 Jun 2011 02:43:59 +0000

[...] different bits of content, we’ll set their course category differently, which is then used for Test & Target purposes…but works well here [...]

By: Tim Elleston

Tim Elleston — Sat, 04 Jun 2011 02:23:10 +0000

Hi Dan
Always happy to share. I’ll post an update when we have some solid numbers. A couple of years ago we ran this as a straightfoward A/B test across one single campaign. The results were really interesting. While we didn’t notice a specific uplift in numbers from the A/B test (we were expecting one or the other to win), what we did notice was that people who experienced only one side of the test (experience B through a microsite) had a higher propensity to complete their application – in fact, 28% higher. That alone accounted for significant revenue.

Then last year we ran a re-engagement test and re-engaged 34% of lapsed applications, which again was significant revenue.

What we’ve done with this one is to combine them all (and chuck in two campaigns), so the results will a) hopefully be interesting and b) tricky to analyse I’m sure

But I will definitely post an update when we get some results.

Thanks for your comment though.
Tim

By: Dan Grainger

Dan Grainger — Fri, 03 Jun 2011 11:15:48 +0000

Great post Tim, coming at a perfect time for me….we’re looking into purchasin T&T in my current company, while the company I’m moving on to in the near future already has it. This is exaclty the kind of stuff I’ll be looking into.

Fully appreciate you won’t be able to release exact figures, but would you be able to give an indication of the success of this when you get further into the test? This is a far more complex test than many people would dare attempt, so I’d like to get an understanding of the incremental value this kind of testing delivers compared to say a more straightforward A/B test.