Experimentation is important, and not testing can lead to some disastrous outcomes for the performance of your website.
This is something I know, and lot’s of other people do to. If you’re reading this, then you may not. Or you do, but you don’t know how to convince your colleagues that they should feel the same.
A podcast I listened to recently suggested that experimentation is a political game, and the more I speak to people about why they aren’t testing, the more apparent this becomes. If you don’t play the game correctly, you are never going to be able to fully build an experimentation culture in your organisation and will find yourself ploughing a lonely furrow. But what are some of the best ways to win friends and show the importance of a testing culture?
The podcast mentioned above, along with some other opinions in the industry, have helped to inform the below steps that can help to grow the influence of experimentation in your organisation.
1. Speak their language
Those of us in experimentation, by nature, work in numbers.
We live for statistical significance and MDEs.
However, not everyone is the same. That’s why you need to tailor your message for your colleagues’ preferences.
If you are speaking with a number’s person, go wild and talk about The Chance To Beat Control, Type 1 Errors, and why an experiment didn’t make the grade despite a noticeable increase in the KPIs.
However, if you are speaking with someone who is right-brained and more creative than analytical, adjust your message to fit with their needs. If their eyes glaze over at the very mention of statistical significance, explain in layman’s terms and find the best way to get the message across.
2. Understand their concerns
Much like a politician needs to press the flesh to be elected, you too need to meet your public.
It’s a great idea to meet each team in your organisation and understand what makes them tick. What are the KPIs they are judged on, and how can you help them to deliver?
This serves a dual purpose of helping you to create a metric tree for your organisation and understand what to optimise, but teams will also be more willing to work with you if they can see that you are helping them to look good and achieve their goals.
3. Have friends in high places
Politics is Power.
As much as being likeable and collaborative can be useful, sometimes you need some muscle.
Many an experimentation blocker can be removed by a command from on-high.
Getting buy-in from Senior Management and the C-Suite can mean that, when all roads seem blocked, Senior Management can insist that testing is performed and have the authority to make the final call on a project.
4. Be Transparent
Of course, this is a point of failure for 99.9% of politicians, but you should be transparent about your successes and failures.
Communication is key, and a weekly or monthly newsletter with experiment results, shout outs to colleagues who inspired the winners, and clarity on where tests failed or were inconclusive (and what you can learn from them) is invaluable.
This transparency and show of vulnerability when things haven’t gone as you had hoped, are key to humanising you in the organisation and getting people bought in to the experimentation programme.
5. Be Likeable
We’ve all worked with someone who is ‘difficult’.
Now think about approaching that person with an idea, or a request for help – would you do it?
We all have our standards in experimentation, our strong opinions around how things should be done. However, don’t cross the line from holding those standards to being difficult and hard to work with.
People across your organisation are going to be more likely to approach you with ideas and requests to run tests, or a desire to learn, if you are approachable and helpful. Although it sounds condescending, sometimes we have to swallow our pride a little bit in the beginning.
I know I said I had 5 steps, but here’s a bonus one that is invaluable for the politics of experimentation...
6. Never say ‘I told you so’
Sometimes people won’t listen to your warnings no matter what you say.
They’ll run an experiment with no backing in user research. They’ll just ship a feature without testing.
On these occasions, when something goes wrong, don’t say “I told you so”.
Be helpful and offer to assist them in avoiding the same issues next time.
No-one likes a smart-arse.
If you’re going to be a politician, be one of the good ones...
Politicians come in all shapes and sizes, so the way you tailor this is key. Showing the value of experimentation is fundamentally the most important thing you can do, and the steps above should help you approach this. And make sure you keep your seat in the next election!
