Kevin Hale - How to Improve Conversion Rates

So this presentation on improving conversion rates is designed to mostly focus on landing pages. But all of the principles and ideas that I'll talk about in this talk actually can help you improve the conversion rates of almost anything, any user interface. So keep that in mind.

This is a typical example conversion rate funnel. When you're trying to improve one conversion rate, you're basically trying to improve the efficiency of going from one step to the next. The thing is, why we care about conversion rate is because it's part of two different aspects of growth, the two main sort of drivers. Growth is kind of like the balance between conversion and churn, and basically, growth happens as a gap between the two.

Something to keep in mind is that working on churn is actually much easier than working on conversion. So in this talk, we're working on the harder thing. It's the thing that usually are going to get started. When I talk to a company and I'm trying to help them with their conversion rates, the first thing I usually try to figure out is do we even need to be working on this at all?

The only time you should be working on and improving your conversion rate is because you have a leaky bucket. So I'd like to just talk about a couple of benchmarks into the industry so that we've all kind of sit on the same page and understand whether we should be working on this or working on something else, like putting more things into the top of the funnel.

Here, conversion rates are about 0.5%. Basically, this is old-school, like before the internet, if people just released software out for free and they just hoped that somebody would pay for it or have their own goodwill. This is the conversion rate you can expect.

Casual download games have a conversion rate of about 2%, so stuff that you sort of play on and off while waiting in line. Most companies kind of care about this one. Freemium software as a service companies range between 1.5% and 5%, with an average of about 3%.

So, once I talked to a company, and they have pages that are converting at about 3%, which means that out of 100 visitors, two to three percent sign up. I used to say you probably don't need to spend that much more time on this; probably other things you should work on instead.

That being said, you can do much better than that. Flickr, back in the heyday, had a conversion rate between five and ten percent. Adult Friend Finder, depending on what you're selling, had people wanting a whole lot more, so ten to twenty-two percent for sex and end of loneliness. Even better are children's social networks, and so basically this is the conversion rate to get your kid to shut the up, right? To leave you alone.

TurboTax Online boasts a monster seventy percent conversion rate. Basically, you're going to TurboTax, you're downloading that software, and you are paying for it; it is very, very high intent.

Every conversion rate problem looks like every other user interface problem, and the concept or framework that I'd like to use to explain how to solve any user interface problem is using something called the knowledge spectrum. This was created by an amazing interface designer named Jared Spool.

Basically, the knowledge spectrum says that this represents all knowledge on a spectrum. On this side represents zero knowledge, no knowledge; you don't know anything. On the other side is God-like, all knowledge. Your product and your user sit on two points on that line, just two dimensions. Your user sits here at what we call the current knowledge point, and your interface, your landing page, the thing you want them to do, is here at the target knowledge point.

Every interface problem that's trying to be solved is trying to close what we call the knowledge gap. That's it. You don't need to go to a complicated design school to know how to solve these problems. You are either going to increase the amount of knowledge that is needed by your user, or you need to decrease the amount of knowledge needed to use your product or interface. That's it. And I, whenever...