Best Multivariate Tester?

Mike

New member
Jun 27, 2006
6,777
116
0
51
On the firing line
From your own experience, what's the best multivariate tester?

I have found several, and the pricing is all over the map, but without prior experience with any of them, I'm loath to spend a bunch of money testing the testers.

I used to use Website Optimizer, but Google rolled it into Analytics and it appears that they removed the multivariate portion of it. Now, you can only do A / B testing up to 6 pages. Well, I don't need Goog to do that.

So, back to the original question: Multivariate testing platforms. Which one's the best.

Must not slow page speed down, and should be fairly intuitive to use.
 


Same here, I've been looking for a simple self hosted solution for a/b and multivariate split testing, with pixel conversion tracking.

I'm sure a lot of people would want something like this, if there isn't anything good available maybe we could have a group buy here on WF to get someone to build it?

Google website optimiser was so good, I don't know why they got rid of it. Content experiments is weak.
 
DIY Landing Page Tool and A/B Testing

30 Day free trial, starting at $49/mo, but you're going to want at least the $99/mo. package.

I worked with Unbounce platform to design landing pages and I have to say, if you are looking to design complex stuff, forget it. Other than that, it's an awesome platform with cool options especially for testing, good user interface and experience and the ability to easily integrate scripts, send reports, data, mailchimp integration etc.
 
I have used all the major AB testing software. I rolled my own many years back and used to give it away to earn clients (splittesting [.] com), but the software became too difficult to keep up since I preferred to not program all day long. Nothing left on that site - but I have plans...

Visual Website Optimizer, IMO, is the best for the money. I also use Adobe Test and Target (formerly Omniture) daily.

What I have found with multivariate testing is that true conversion (sales) results often don't hold up over time. If you do multivariate testing - you better take the winners and then do AB testing with just them. I basically use multivariate testing to quickly determine a small group of winners that I will AB test. You need a lot of traffic for multivariate to work out.

Also, don't make decisions on big changes unless you have both a lot of users in the test - and time. Most of the AB testing packages will declare a statistical winner after a predetermined threshold has been hit (I think Test&Target uses Student T-Test )

Don't just go off of what the test says is the winner - use common sense and especially use *time*. I have seen countless tests quickly declared statistical winners only to have them bomb out when left to run over 2 or 3 weeks.

Always test.
 
Same here, I've been looking for a simple self hosted solution for a/b and multivariate split testing, with pixel conversion tracking.

I'm sure a lot of people would want something like this, if there isn't anything good available maybe we could have a group buy here on WF to get someone to build it?

Google website optimiser was so good, I don't know why they got rid of it. Content experiments is weak.

The reason no one builds a self hosted version is that this industry is full of god damn thieves. From a profit standpoint building a SaaS like optimizely or unbounce is much much better.
 
The reason no one builds a self hosted version is that this industry is full of god damn thieves. From a profit standpoint building a SaaS like optimizely or unbounce is much much better.

The only way you'll have a self hosted version is if you hire a competent coder to build it or you code it yourself. It can certainly be done. I successfully had my own self hosted CRM coded and use it for all my offers.
 
The only way you'll have a self hosted version is if you hire a competent coder to build it or you code it yourself. It can certainly be done. I successfully had my own self hosted CRM coded and use it for all my offers.
How much did it cost you? And now that you have it is there any reason why you wouldn't make a business out of selling it?
 
How much did it cost you? And now that you have it is there any reason why you wouldn't make a business out of selling it?

I might offer it as a SaaS. The response was kinda lukewarm in the Skype chat room I was in. Nobody wanted to switch from Limelight CRM to an "untested" offshore hosted CRM. They said they wouldn't trust an offshore CRM with their consumer data, even though I'm not sure what makes Lightlight so trustworthy since they're in the USA. All it's going to take is a civil demand from the FTC...

I might consider offering it down the road though. I'll have to add in some modules to make it easy for clients to switch over from other CRMs.

I won't be selling it as a self hosted solution because it'll just be a great way for people to rip me off.
 
seen this the other day here, someone posted it. Forgot who. + rep to whoever it was.

I thnk its a/b only though.. look at it

https://github.com/gregdingle/genetify/wiki


+1 for genetify. It's free. It's fast. You can self host. The big BUT is that is that you need to be able to place code on the purchase page for a true measure of conv %. Any thing else (eg LP code) will just get you what combos have the best CTR.
 
I'm the creator of Landing Page Genius, so I thought I'd chime in here.

LPG is self hosted and works with either Prosper202 or CPV Lab. All the data it gathers is sent to either one of these trackers.

It's not dumbed down and point and click like some of the more robust SaaS solutions, but it also isn't going to cost you hundreds or thousands of dollars a year.

Genetify is great for what it is... but beware... anyone with a web browser can see all the results of your tests. There is no security whatsoever.

If you guys have any questions about LPG feel free to post them here and I'll answer them.

Cheers