Optimizely Says Good Decisions Are Not About Hippos

Optimizely Says Good Decisions Are Not About Hippos
Facebook may have decided that you shouldn’t see the news, but we think you deserve to be in the know with Lifehacker Australia’s content. To sign up for our daily newsletter covering the latest news, hacks and reviews, head HERE. For a running feed of all our stories, follow us on Twitter HERE. Or you can bookmark the Lifehacker Australia homepage to visit whenever you need a fix.

Innovation is not easy. It involves experimentation, taking risks and investing resources in projects that may not deliver the outcomes you want. Optimizely takes a data driven approach to testing multiple scenarios so you can make better decisions about what’s best for your business. I spoke with their new managing director for ANZ, Dan Ross, about their Australian launch and the opening of Australian offices in Melbourne and Sydney and how decision making is not about hippos.

Optimizley was founded by Dan Siroker, a former Google Chrome product manager and Pete Koomen, the product manager for Google’s App Engine. Siroker went on to work on the presidential campaign for Barack Obama. The challenge was finding ways to convert contacts into political donations – the lifeblood of a presidential campaign. The lack of good tools for doing A/B or split testing led him to launch Optimizely.

Ross said “The existing technology was not sufficient. We’ve now built a business on the power of experimentation and scientific method for better interacting with customers no matter where they are, whether that’s website, mobile, call centre or survey results”.

You need to go further than traditional data-driven decision making. That relied on conducting lots of analysis on existing data. But the result was often more questions than answers. That’s why experimentation is valuable. It lets you see the consequences of different decisions, or non-decisions, in order to iterate and refine your actions.

Depending on the “hippo” – highest paid person’s opinion – or the knowledge and experience of a small number of people is not good enough. It’s simply not possible for any one or a small group to know 100% of everything all the time in a complex business.

Although the company has, until now, not had a strong local presence they have made inroads in a number of different industry verticals with Atlassian, Optus, AGL Energy, Chemist Warehouse and Fox Sports already on their books. Ross notes Amazon’s coming entry to local Australian markets is seeing businesses look increasingly at tools like Optimizely to enhance their existing services so that they aren’t swamped next year.

“Amazon has done exceptionally well and has a deeply ingrained culture of experimentation. They take a hypothesis, test it, learn from it, and iterate at scale,” said Ross. “We think we have the technology to allow a business that didn’t grow up like Amazon, Google or Netflix to do this in a competitive way”.

Quantifying the benefits of these experiments can challenging. Ross talked about Booktopia, which acquired Angus and Robertson back in 2015, but wanted to find ways to leverage Angus and Robertsons’ customer base.

“They formed a number of hypotheses,” said Ross. “In a short amount of time they doubled the revenue from that side of the business. Using experimentation, they figured out what would resonate with that group of customers and tailored the experience from those learnings”.

Another example came from the energy sector. A large ASX-listed retailer was looking to enter a market where they were not well-known. The management team, leaning on their experience, formed six different variations of the market entry messaging.

“They ran a test using Optimizely and found one of the six performed way better than the others. It was not the one the executives expected to work best,” said Ross.

Looking at those examples, the key take home lessons are that management by experience and gut feel are no longer adequate, and it is possible to gather sufficient data if you use the right tools for conducting experiments.

“We have the data. We just aren’t making great use of it,” said Ross. “The challenge is do this at scale is hard We think we have the technology to do this”.