Alternatives to GWO (Google Website Optimizer)? [closed]

We're using GWO (Google Website Optimizer) now. The multivariate and A/B testing is exactly what we need and works great from the perspective of showing the variations to the users. However, we have several issues that make me want to use a different tool:

  • Statistics are inaccurate compared to Google Analytics, so we now disregard them and have to manually check
  • Previews typically don't work
  • Cannot have dynamic content in variations (I know about variation_content, but I cannot get it to work and nobody in google's forums has been able to help.. I suspect google may have stopped supporting this)
  • Documentation is poor, there's a techie guide with well-known inaccuracies which haven't been fixed in well over a year.
  • The html/javascript code we modify our multivariate test sections with is ugly and makes our pages fail standards validation
  • Only 8 test sections per page, problem there is we want to allow our marketeers the ability to do everything they need from within GWO, but now they need to mark off which test sections they want/don't want in our custom tool
  • Different experiment key for every test, again it makes marketeers need to work with our code sometimes

Is there a good tool like GWO that works with Google Analytics (which I love)?

UPDATE: We went with Optimizely and have generally been happy. However, it can be difficult to work with because it does a little too much for you. You edit your webpage directly from their UI, but of course that isn't always easy or even possible. Particularly when Javascript is involved. Our UI often gets screwed up in the process. I liked GWO's approach to this in that the developer sections off the code and the marketer can then fill in those sections with variables the developer allows for. To me that's ideal, except that GWO, of course, doesn't actually work.

There's a very similar competitor to Optimizely called Visual Website Optimizer. Also looks very nice, but has the same issue I describe above.

Is there a GWO that works?

You should take a look at Optimizely.

  • Doesn't require creating invalid code.
  • Easy to create variations on the fly, though only A/B, not MVT.
  • Simple WYSIWYG test design, on the fly.
  • Real time data.
  • Retroactive goals
  • With regex/head matching for experiments, you can set the experiment to work on dynamic pages.
  • You can set a Google Analytics custom variable for the experiment that will pass the variation the end user sees as a custom variable. (It even allows you to set what slot you want it to use.)
  • The test variations are basically just jQuery manipulations of the DOM, so if you know a little jQuery, it's very easy to extend its capabilties even further than the very generous WYSIWYG GUI.
  • Installation is easy: You only need to include a single script tag, one time, on any experiment or goal page.

I have found Adobe Test&Target to have all the features you need. It is very easy to create experiments, add variations, and conversion goals. You can easily inject JQuery snippets to create new variations, click Save, and the test is running in production.

I have no idea how much it costs, but I'm guessing it is not cheap.

Now Google Website Optimizer killed multivariate testing in new version (Google Analytics Content Experiments) we launched Convert Experiments on for people that look for a GWO alternative with MVT

Yes I am Founder of Convert Insights, the company behind this tool... Dennis

Re your update: I have tried both GWO and Optimizely, and I'd go with Optimizely every time.

You say you wish it worked a little more like GWO - if you want, instead of manipulating the elements of the design using their GUI, you can just redirect each variant to a different URL:

There are a few other tools which do A/B and MVT. Aren't free, but check them out for yourself: Omniture, Webtrends Optimize, SiteSpect.

Hope this helps!

You can also try VWO. It does MVT as well as A/B testing and is also a good tool. Optimizely is easier to use though so you might want to evaluate both for your scenario.