Some notes on the conclusion of Google Optimize, a thing that we all saw coming
Google are not a trustworthy company.
For two decades, they have built products that boldly enter industries, used its advertising business to prop up margins internally, and then exited when they aren’t making enough money to support those independent products. This happened with project management tools, RSS readers, and chat. If you wish for more details, you may visit the graveyard.
So when Google announced that Optimize, its once-beleaguered online experimentation tool, was getting a serious upgrade in 2016, I was skeptical. I put a lot of effort into understanding the ins & outs of an experimentation framework – and at the time, I was all-in on a competitor. My A/B testing framework is a load-bearing component of my job. I didn’t want to become dependent on Google to do my job, only to have them pull the rug.
At the same time, I looked at the competition. My platform of choice, VWO, had not updated their JavaScript tech stack in the five years I’d been using it. (They still have not, forcing an incorrect sideloading of jQuery 1.x on every customer session.) They hiked their pricing, limited my agency account, and started to strong-arm me into onerous long-term contracts. I looked at Optimizely and discovered an even worse situation, with “call us” pricing and five-figure enterprise-level contracts.
There is still a place for experimentation for small & mid-sized businesses. The addressable market is sizable. Experts like myself can help show others the way. After all, most of our work is in finding what to test and how to test it. That exists independent of any framework or tech stack.
So I’m not surprised to see that the industry did this to themselves. Google Optimize’s dominance & death is the single biggest own goal in the history of value-based design. I saw it coming from day one, and it still happened, and we only have ourselves to blame.
As an industry, value-based design remains nascent. Tooling is not evenly distributed. But eventually, WYSIWYG DOM parsing and per-variant analysis will be baked in everywhere. Heck, someday we might even see native experimentation in Shopify. But for now, we’re in a fallow period. Not enough tools exist, and the ones that do exist charge too much, don’t do enough, and do it all the wrong ways.
Quite frankly, VWO needs to stop playing like their aging framework is acceptable. Optimizely needs to support the long tail now that their funding has run out. And everybody else needs to begin posting actual pricing to their public-facing pages. And that pricing should actually support the ongoing cashflow needs of small & mid-sized businesses who can still strongly leverage experimentation.
I shouldn’t have to tell a $7M store that their tool will cost $12,000 a year on top of the human that powers that tool. I want to be able to look at how much something will cost and buy an account without having to talk to a person. I can’t believe I have to say any of this out loud.
Whomever makes this happen will get Draft’s money for many years into the future. As things stand, Convert and Omniconvert do this, and I remain optimistic that others will get the memo as well.
I also suspect that Microsoft will see this as an opportunity and expand the already-excellent Clarity into optimization. As for Google’s promise to incorporate optimization into GA4, I’ll frankly wait until I see it. As-is, their announcement is profoundly fear-inducing without a working backup plan, especially coming a week after mass layoffs were announced for no good reason.
For specific tasks, there are quite a few options:
- For landing pages, there is Shogun, Zipify Pages (a former Draft client!), and Justuno.
- For pricing, there is Intelligems.
- For email campaigns, there is Klaviyo. I guess other email marketing platforms exist? Whatever.
We have 9 months to adapt to a new normal. In the meantime, if Google gives you another tool, will you trust them, knowing what they’ve done in the past?
This week, for paid members
- In our next Store Design lesson on research, we’ve written about deep dive surveys. How do you build, critique, and analyze them?
- Our design of the week shows a page that looks shockingly uncredible at first blush – but it works. How?
Want in? Join us now – named one of the best ecommerce communities going on the web.
Links
- Essential data dump on Shopify, from how many stores are operating to how much revenue they generate.
- HTML has updated their specification on the
<dialog>element, making it viable for broader use. - 21 thoughts on information architecture. Question to think about: what is the relationship between information architecture & value-based design?