We mourn Google Optimize, which will soon be no longer available

TRANSCRIPTION OF THIS EPISODE OF THE PODCAST

My name is Jörg Dennis Krüger, and, as my PC technician at reception has already said: 

Yes, I am the conversion hacker. 

And in today's edition of the Conversion Hacking Podcast, I would like to mourn a little. Unfortunately, Google is not only shutting down Universal Analytics, which is bad enough, because everything is different in Google Analytics 4, some things are a bit more complicated and so on. But much worse: Google is shutting down Google Optimize without offering us an alternative. 

The reasons for this are varied or not entirely transparent. We can only speculate, but it seems as if Google simply does not want to compete with commercial providers. And it also seems as if Google has invested in one or the other commercial provider and simply does not want to put any more resources into its own development of Google Optimize. 

This is a great pity, because Google Optimize was or still is a really good A/B testing tool. With Google Optimize, we were able to do everything we wanted to do in testing, at least almost everything. 

We were able to do excellent personalization, and the whole thing was completely free and fully integrated into Google Analytics. That's a really great thing, of course, and yes, where are we now? What would we do without Google Optimize? 

On the one hand, we can continue to use Google Optimize for the time being, because Google Optimize is still there temporarily and is also integrated into Google Analytics 4, which came as a major update last year. But we will no longer be able to use it after that. That means we have to see what tools are available on the market. 

There are the big and perhaps once very well-known ones like Optimizely. But that is expensive. It can do a lot, but it has become an absolute enterprise solution. Who else do we have, especially in the European market, when you think about data protection issues? 

We have A/B-Tasty, which is sometimes cheaper, but is nowhere near as open and flexible as Google Optimize. There is Camelion, which is significantly more expensive. You still spend a lot, it's a very sophisticated tool, but you really have to have a proper A/B testing team and A/B testing concept to make the investment worthwhile. 

And then there's Web Trends Optimizer, which is no longer part of Web Trends. A few years ago, Web Trends Optimizer was sold out of Web Trends. There is Web Trends Optimizer, they're based in the UK, and it's not a bad tool at all. I use it too, sometimes for A/B testing, I think it's pretty cool and it can do a lot. However, it has a slightly different logic than Google Optimize, for example. You have to get to grips with it first, and the reporting isn't that good, but it can do an incredible amount and isn't that expensive. 

So, you can get a nice solution for a bit of money. The entry-level solution for small projects is about the same price as Camelion, but it scales up better in terms of price. 

So who else do we have? Sure, there's Adobe, the test target, the system I started with, with A/B testing. 

Then there's convert.com, for example. I think it's a wonderful system and very easy to use. However, the editor is not quite so great. But most tools really have problems with that. Google Optimize is also incredibly far ahead with its editor. Maybe the editors will get a little better. 

Most people now work with plug-ins, especially for Chrome, in order to create something better. But "Convert" is relatively cheap, with 99 € per month you can already do optimal things. And yes, data protection, they are based in the USA, but if you integrate them properly (and they don't collect IP addresses, you can switch that off), it should also be fully compliant with data protection regulations.

I think this whole data protection effort is a bit exaggerated anyway. As long as you stick to certain basic rules, such as anonymization, which is the most important basic rule, you're almost always on the safe side. This is not legal advice, but this anticipatory obedience to really map everything 1000 percent, you have to take a quick look at whether you are actually big enough for that. 

Of course, if you're Zalando or something like that, then you should pay proper attention to data protection and make sure that everything is really 100 percent. Otherwise you can get into real trouble. On the other hand, if Zalando were to be fined 50 million euros for not complying with data protection regulations, we could also discuss whether the investment might be worthwhile. 

(But that's not what I said.) 

That's why you really have to look at it, and the big companies in particular think this way, and many data protection officers overshoot the mark. So, if you use a solution here that simply works anonymously, you're well on your way. 

Another classic solution is Visual Website Optimizer. It has also become quite expensive, they all used to be much cheaper, but can do an incredible amount. They have strong connections in India. In principle, it's an excellent tool because they can put a lot of development work into it for the money they earn from it. However, the support also goes through India. And because of the data storage, you have to take a close look. So it's also a good tool. They can now do an incredible amount, but I think it's a bit expensive. So I wouldn't necessarily use it, of course it depends a lot on the project. 

Yes, well, we can't do proper A/B testing without actually spending money. Of course, there are a few ideas and scripts that you can use to put something together yourself. You can use a tech manager to put something together for free and then run an A/B testing variant on it, track it in Google Analytics and then evaluate it. But it's a pretty complex solution.

So, you lose a lot of the convenience and a lot of the possibilities. Starting with proper staging and URLs that you can share for approval, proper and regular reports and so on. That's why I would only use it if nothing else really works. I started with Omniture around 2006 and wrote my book in 2011, and in between I also built something like this myself, which in principle is relatively simple. 

When calling, you set a cookie and then have a table or a Rae where you roll a random number and deliver the variant. Then you write in the cookie which variant you have delivered so that you can deliver the same one again. The variant is just Java Script and then you just send a hit to Google Analytics and track it accordingly. So, that works. But you have to program everything, every change. When it comes to precise work such as targeting and so on (i.e. targeting certain visitors who only come via mobile or desktop, or from certain sources, etc.), it becomes increasingly complex. 

Sure, you can program your own A/B testing tool, you can also build your own car. It's all possible and it's all cheaper at first glance, but then you see whether it's really even cheaper with the work involved and so on. 

So, what am I trying to say? I am very sad that we have now lost Google Analytics. But Google is unfortunately not a reliable partner. You always have to realize that Google does funny things. They've also just introduced a limit on the number of files on Google Drive, even for paying customers. And then they removed it again after a few days because the customers all ran amok. Because when I had a lot of small files on Google Drive, even if I was far away from my, I don't know how many terabytes of storage I had rented, I suddenly got an error message saying that my storage was full. That doesn't work at all! Google has also done away with a lot of tools in other areas, and we know that, we used to work with Google Plus and so on ages ago, and then Google just let it die. 

So, Google is a bit of an unreliable partner. In this respect, it is perhaps also relatively good to rely on a commercial solution for a proper testing program, to commit to a partner that is really there for the long term. You can see it with Google for the customers of Analytics365, where Google now has all the tools available for even longer. Google Analytics Universal, just like Google Optimize, (although it is expensive, you have to know that). But yes, finding a reliable partner is certainly not a bad thing. 

And it is essential that good A/B testing does not depend on the tool. With a €99 convert.com, we can probably achieve exactly the same test results at the beginning as with, say, a €1500 Camelion or VWO, or whatever Adobe now uses. So yes, just start testing and of course integration is a bit more work now, but it's not really that much. 

And if you want to be taken in hand, well, you can of course knock on our door. We have the Think Conversion Masterclass, which is all about A/B testing. Of course, we've already done quite a bit of A/B testing here, and I in particular have done it in recent years. 

Well, in any case, keep in mind that Google Optimize will soon no longer exist. Use a commercial solution, do really good A/B testing, really well structured, with a proper long-term plan, and yes, then A/B testing will also work without Google Optimize. 

That's it from me. Thank you very much until the next issue. There will be more issues now. I hope you've already seen the new rhythm. We're really stepping on the gas now. My team has also grown a lot, so we now have a bit more resources to do this properly. And then we can get going and disseminate a lot more information here in the podcast. 

Also the reference to Instagram, Jörg Dennis Krüger. We regularly post information there, we regularly post stories with real added value and follow my personal profile on LinkedIn or the JDKRÜGER&CO. company profile. More will happen now, because we want to deliver knowledge. 

We want to ensure that all online stores in Germany, Austria and Switzerland become better, with our collaboration or at least with our knowledge of our social media channels. And if you have an English online store or are active in the English sector, spoiler alert: The JDKRÜGER Corporation in the USA has already been in existence for some time, as some of you may have seen. This is now being brought to life and we will soon be offering our entire service in English. So, if you're interested, just let us know. 

Thank you very much, I wish you many conversions, and until the next edition of the Conversion Hacker Podcast, see you soon!