A list of your own problems
How do you optimize an optimization effort?
You have 12 months to get a 20% increase in your conversion-to-paid numbers. How do you do it?
First stop, Google. You search for conversion optimization tactics, a list of the latest hacks, and you end up with a bunch of articles. Let’s say you manage to compile a list of 100 techniques to improve your conversion rate.
Which one do you implement first?
You could try and implement them all at the same time but then everything would cancel each other out and you’d have no understanding of what worked and what didn’t.
The alternative is to test each technique in sequence. Let’s say you have enough traffic that you only have to run each test for a month. It will take you 8 years to work through all the tests.
The problem with growth hack listicles and articles about product development is that there’s no way to prioritize what you need to work on first.
Then you have best practices, benchmarks and industry leaders. They’re a good place to start but best practices are also common practices. It’s what everyone else is doing. How do you beat the competition when you’re all following the same playbook?
So the question remains, “What should we test next?”
If your testing program and product development process is based on guesswork and a bunch of random hacks and best practices then you need a more systematic way of finding opportunities for growth opportunities.
So what does a good optimization process look like?
The first thing it does is tell you where your problems are.
A good process establishes an understanding of where the biggest leaks are and a clear, shared understanding of why these problems are even problems, to begin with.
Once you can identify and rank problems then you need to be able to come up with lots of hypotheses for ways you can fix these problems. The more people these ideas come from and the more real-world evidence they are based on, the better your process is.
Finally, you need a way to prioritize your ideas for solutions to the problems you’ve identified.
A good optimization process results in testing effective changes and this boils down to understanding what matters.
If we actually know what matters to the people who use your product, what they like, the kind of pain they go through, and what emotional need you’re helping them satisfy, you can ship a product and message that resonates.
You can’t highlight features they care about or talk about stuff that matters to them till you actually know.
If you don’t know, you can’t optimize.
So the discovery of what matters is the crux of any effective optimization program.
So how do you figure out what the real problems are?
There are two ways to do this. The first is by testing stuff to see what works. One is by testing. The second is by doing the legwork of reserach. You should be doing both.
Research is the process of empirically answering important questions you have about your product and the way people use it. Research can be based on quantitative data, qualitative interviews, screen recording, surveys, etc. It’s usually a combination of technical analysis, behavioral analytics, heuristic analytics, qualitative inquiry (like surveys and interviews), and user testing.
What is important is that you establish a list of your own problems so that you can systematically uncover what actually matters in relation to your specific product and unique audience.