Neil Patel of KISSMetrics on Successful Web App Analytics
By Keir Whitaker @keirwhitaker
29 April 2010 | Category: Web Apps
At Future of Web Apps Miami 2010 Neil Patel of KISSmetrics gave the audience 5 "must measure web app metrics". It's a short session but full of great hints and tips from someone who really understands this area. If you haven't checked out Neil's blog I really recommend you do.
Watch this and other videos directly on our Vimeo Channel
Transcript
Ryan: We work quite hard to pick topics that we think are relevant for you guys, and one thing that I think is really hard when you run a web app is analytics. It’s overwhelming to understand what you should be measuring, to see if your web app is succeeding; there are a ton of packages and it’s quite confusing.
We think that as developers you are able to build your apps just fine but when it comes to the marketing aspect, it’s quite confusing. We asked Neil Patel to talk about analytics and 5 things that you need to measure to figure out if you’re succeeding or not. I think it’s going to be a very useful session for you. Again, if you have questions, don’t feel afraid to ask. Please welcome up Neil.
Neil: Hey everyone, as everyone said my name is Neil Patel. I’m with a company called KISSmetrics. Today, I’m going to be giving a presentation on the 5 metrics you ought to be tracking. It’s pretty simple; there are 5 main points that are easy to track and easy to take notes. Here we go.
1. Revenue Per User
A lot of you track conversions. Am I correct here? If I told you that I could increase your conversions from 5% to 10% you would be ecstatic, correct? In most cases, if I increase it from 5% to 10% you’d be like, . That’s great. That’s going to be a lot more money.” On the flip side, if I said I could decrease your conversion rate from 10% to 5% and make you triple or quadruple the amount of money, you’d be like, . What?”
A lot of times when you’re on your web app and tracking metrics, what you have to track is conversion per user. An example of this is we have a SaaS based product tracking how much revenue is coming from each user per page. We use Google Analytics to do that. One thing we tested was we were charging $97 a month. I took that $97 a month price point, and I said, . What if I charged $197? What will happen?” Amazing, by roughly doubling the price, the conversions decreased by 10%. The cool part, the revenue increased by an overall 82%. That was a pretty big jump. We got a decrease in conversion but it was still worth it because we made more money out of it.
If you want to track your revenue per user, you can go to Google Analytics. A lot of the things I’m going to mention are using Google Analytics because it’s free; everyone can use it. Sign into your account, click . edit” next to your profile, go into the settings, next into the website information, and find the . ecommerce” button. Change it from saying no to yes, and you can start tracking quite a bit of information.
They’ll give you some code, and some of the things, if you can see here, transaction, order number, total amount, tax, and so forth. It’s the same thing with the item number; item number, what it is, color, all that kind of stuff so you can actually track the revenue per user. I’m not saying you shouldn’t track conversion rates, but you also need to track revenue per user so you can get a better understanding of – are the changes you’re making on your website not necessarily just increasing your conversion or decreasing, but is it making you more money or not.
2. Engagement
With a lot of web apps out there, if you have more engaged users, you’re more likely to get them to pay, invite their friends, be stickier, and so forth. A really cool metric is how can you get really engaged users. It’s one thing to have a thousand or a million users on your site, but I would take a hundred really engaged users over ten thousand users who come to the site once and never come back and they’re not spending money or making the app grow.
Some of the things that I like to track with engagement is when someone signs up for my app, let’s say it’s facebook.com; I would go out and say are they signing up? Once they sign up, what else are they doing? Are they entering their email, profile name, adding a picture, putting in descriptions, etc? How far in the process are they getting, overall?
In addition to that, you want to see how often users are coming back and what is causing them to come back. In some cases, it could be that email notification the Facebook app is sending saying, . Your friend did blah, blah, blah.” Or, it could be something else. It could be that they’re coming from a search engine, or whatever they may be.
You also want to track what your users are doing. There are a lot of features in an application. The ones that are really engaged and sticky, are they coming back and doing one thing, such as playing a game on your social app, or is it that they’re using a specific feature set, such as if you’re creating a form, or looking at the stats. You have to figure out what is causing users that are engaged versus what is causing users that aren’t engaged. The ones that are more engaged are more likely to pay for your product, talk about it in a positive fashion, and tell others to sign up.
A lot of times, with engagement, if you’re using something like Google Analytics, you need to be tracking your internal site. What happens is a lot of people are like, I got my About page, I got my Contact page, I got my Home page. This is all the traffic I’m getting.” How much traffic is going to the dashboard? What are people doing? Google gives you some code and you can put it on internal pages and see how much traffic these pages are getting.
3. Event Tracking
How many here know what event tracking is? Google Analytics and a lot of analytic solutions out there give you this thing where you can track specific events on your site. It’s the best way to find out if a feature set in your app is working for you or it isn’t working out for you.
Why would you want to invest a lot of time and money in development; most of you already know that development takes forever. So you don’t want to waste money in features that aren’t really being used or that people don’t really care about.
One thing we tested out was the chat feature. How many people are actually using this chat feature, how engaged are they, and from there we figured out whether it was worth keeping or worth removing. If it was worth keeping, should we invest more time and money into it?
What ends up happening is that a lot of times you’ll get quite a few people using a specific feature, and they’re using it tons of times, so the page views rack up. But you’ve got to figure out if it’s just the power users that are using it; is it 90% of the user base that doesn’t understand this feature and doesn’t care about it or even know about it. You really have to track things.
Here are some cool things you can track with event tracking, if you’re not too familiar with it. You can do things like; did someone click play on a video? Did they pause the video? Did they download it? Did they download a PDF? You can literally track almost anything through events and see what people are using.
A good case would be for a video site. Are people using the pause and play button? Are people downloading? That will give you understanding; if no one is downloading your videos and you don’t care, then maybe you shouldn’t have a download button. Focus the user; have them do what you want them to do. The things they aren’t doing, strip them out. That way they’re doing the things that hopefully gets you more engaged users and makes you more money.
4. Real Conversion Rate
Real conversion rate is pretty tricky because a lot of applications out there have a freemium model. What happens if I said my app is mint.com? I first find the site by Googling . saving money” and it takes me to mint.com. What happens after that? I don’t sign up. I end up going to another site a week later, TechCrunch, and I read a story about Mint and how they’re doing all these cool things. I still don’t sign up but I think it’s cool.
A week later, I go to Mashable, and I read another cool story about Mint. Now I’ve seen it three times so now I have to sign up. Technically, most analytic solutions will report this person came from Mashable, they ended up converting. That’s where your conversion point is, and you need to track that, but you also need to track that this person originally came from Google, typing in a query . saving money”. They came a few times from Google, TechCrunch, and then Mashable, and then they converted. You have to look at the original traffic source that’s causing the conversion.
In addition to that, if I have a “freemium model” on my SaaS-based web app, I have to track that someone came from Google, signed up from a free account, then used it for a month and did all these steps, and then they converted into a paid account. Although they’re a freemium user and signed up through Google, you also need to track that this freemium user converted to paid, so this traffic source caused the paid user. With this, I haven’t found a good analytics solution that really does this, so often you have to build it internally to track a lot of this data. It’s worth tracking.
5. Test, Test, Test
A lot of times, and I’m seeing this with a lot of web apps, people are basing their decisions on what they’re reading on the web. A good example for this is 37signals. I love them to death. They’re a great company, smart guys. But, what worked for them may not work for you. Just because you read something on the web doesn’t mean you should copy it.
An example of this is C plans and pricing. I believe Ryan wrote a blog post on how they doubled or tripled their conversion rate, or increased by 100% or something. I forgot the exact number, but after I saw this, there are so many web apps that started copying it. DoItYourself Themes uses it. EasyTweets uses it. The numbers are endless and if I do a search for C plans and pricing, I probably could find a ton more.
The thing is; what works for 37 may not work for everyone else. At my company, we have a web app that is analytics related and it shows you heat maps. We did some surveys on some people and decided to test a lot of different variations out there. We were like, . Alright, C plans and pricing is great, but how can we actually relate the message to what the user is actually doing on our website?” We said, . Show me my heat map.”
We ended up doing a Google website optimizer test, and what we learned is our conversion rate increased by 112%, which is pretty significant, right? That’s a big jump in revenue and it’s not doing something that everyone else is doing. C plans and pricing is great; if it worked for me, I’ll be using it, but it didn’t necessarily work for me. What works for one company may not work for everyone. You have to figure out – you have to create a ton of tests, and if it works for me, great; if it doesn’t it doesn’t.
I use Google website optimizer a lot. It’s pretty simple. You take the original page that you’re trying to create variations of when you do A/B tests. They spit out some code, you put it on the top of the page, at the bottom; you take the variations and then you put some code that they spit out at the bottom of that and then put some code on the conversion page. It’s pretty simple. It looks something like this. There is control script, tracking script, and at the end they’ll tell you.
When you’re doing a lot of tests, some things will work out for you and some things won’t. Don’t let it get you down if a test doesn’t work out. What we’ve learned is if you make a lot of small tweaks and you don’t see a big improvement in conversion, that’s when it’s a good time to try a new design or layout. Don’t be afraid to experiment. Any questions?
Ryan: We’ll do a couple of quick questions.
Q: One of the things we do is event tracking, but we capture so many events that it becomes a hassle to report off of that. Do you have any suggestions?
Neil: What you’re asking is you’re tracking so many events that it’s just becoming a pain in the ass to go through them all to see what’s working and what’s not? You should track all your events, but that is something more as in a time issue.
The ideal thing is to continue tracking all those events and start digging through the data. If you can’t do that, what I would recommend is look at a handful of events, the ones that you think are the most important or the least important, or ones you’re going to be basing future decisions on, for like the next 30 days. Start tracking them, see if they’re working out well, and if they are then based on this say . We’re not going to spend more time on this.” If it’s working well, you’re seeing that it’s causing more engagement, then say, . We’re going to invest more time on this feature and see if we can crank it up even more.”
Use baby steps. That’s the biggest problem with analytics. There are so many pages and so much data that it gets overwhelming to look at everything. You have to take baby steps, concentrate on one thing during the next week, and from there you can rotate and go after different metrics.
Ryan: Cool – so unfortunately we don’t have time for more questions, but Neil is going to be around all day, so please grab him. He’s got a ton of good advice about measurement, analytics, what to do, and what not to do. Please give him a big hand. KISSmetrics is also a pretty cool tool. That’s how we came across Neil. It’s definitely worth checking out.
Follow @thinkvitamin on Twitter Please check out Treehouse
