Kissmetrics Blog

A blog about analytics, marketing and testing

Built to optimize growth. Track, analyze and engage to get more customers.

Customer Analytics: How Analyzing Real People Will Improve Your Business

So you just pulled the report on your monthly revenue. And… PARTY TIME! It’s up by 50%! And since we launched that new marketing campaign last month, it must have worked!

Let’s grab the whole marketing team and have a drink that Don Draper would approve of!

But what if that marketing campaign wasn’t responsible for the increase in revenue? Maybe it was the updated version that the product guys launched. Customer service also rolled out new policies. Do they get any credit?

Here’s the thing: we LOVE taking credit when results are positive but AVOID credit when results are negative. So when business is good, every department claims it was their work that made the difference. Then when business is not so good, it’s always the other guys’ fault.

This exact same situation happened to me a couple of months ago.

I was working with a client last October. They have an ecommerce store that sells brand swag and they were doing about $1,500 a month in revenue. After taking a quick look through the site, I realized that they could make their “add to cart” buttons more prominent, simplify navigation, and change the color scheme to match their brand. I expected these changes to increase conversion rates and monthly revenue. So I launched the improvements by the end of the month.

Want to know much revenue they did in November? $5,580. And in December? $7,240.

So I should get credit for increasing revenue by over 482% percent within 2 months? WOOHOO, gold star for me!

Not exactly.

The rest of their marketing team also released several new products that started to sell very quickly. AND December is the peak of the fabled holiday season.

If we’re only tracking aggregate revenue, all of us want to take credit for the spike in sales (including Santa). With all this noise in our data, it’s very difficult to determine who gets the most credit. Curse you Santa!

So, how do we minimize the noise in our data and find the real signals for improving our business?

We focus on people.

Start With Segmenting

When we only look at the aggregate numbers (even for important things like revenue), it’s impossible to see what’s driving the end-result.

To get around this, we need to break our data down.

The super-official certified analytics word for this is segmentation. It’s the fancy way to describe breaking our data into chunks so we can compare different groups of data.

Let’s say I gave you a report for last month with these metrics for your site:

  • Revenue = $6,930.07
  • Conversion Rate = 3.1%
  • New Accounts = 43

Honestly, this data is kind of worthless. Why? Well, there are two reasons.

First, we have no context for the data. In other words, is a monthly revenue of $6,930.07 good or bad? We have no idea. If the last 6 months averaged $3,000 per month, that’s pretty good. But if the average was $23,000, last month was terrible. This one is an easy fix, compare your metrics to historical data and you’ll get a much better idea of what these numbers mean.

The second problem gets a little trickier. Remember, these numbers are just averages. That means they’re hiding all sorts of trends about our customers. What if 4 customers produce $5,000 worth of the revenue? Or if one of our marketing campaigns had a 20% conversion rate? If we only look at averages, we’ll never see trends like these.

Let’s look at an example of how this works in action.

Here’s a funnel that I’ve segmented by campaign source:

The product gives you an easy way to collect feedback from your users by putting a short survey on your site. The funnel to track how many people sign up for a new account and how many of those people also create a new survey. So the funnel shows us how many people move through each step. This tells us whether or not people are using the product.

Now, I’ve also segmented the funnel by campaign source. This breaks the funnel down into different marketing campaigns so that we can compare them to each other.

What does this report tell us?

We have the average conversion rate which is about 6%. Then if we look at our campaigns, we see many of them are right around that same point. But the “6easyinsightspost” campaign had over DOUBLE the conversion rate, coming in at 13.7%. That’s awesome.

Compared to our other campaign sources, “6easyinsightspost” is performing WAY above expected levels for these conversion rates. Now I want to go look at that campaign and try to figure out what made it so successful. I’ll probably come up with a few ideas. Then it’s my job to put those ideas to the test and see if I can replicate that performance.

Granted, the “6easyinsightspost” campaign only had 19 total conversions which isn’t that many. If we scale up the traffic, I would definitely expect that conversion rate to come down a bit. But even if we see a moderate drop, it’ll still be one of the best campaign sources in this list.

Main Takeaway: Whenever you see an aggregate number, alarm bells should start ringing. You should say to yourself: “I see this data here but what is it hiding? How can I separate this data and break it down so I can see what’s really going on?”

But we can take this concept even further. Instead of breaking up our data it segments, we should also break it into individual people.

Looking at Specific People

We know how powerful segmentation is. But what if we took it a step further and look at individual people? Well, the data gets even better.

Let’s pull some more data on that survey product.

Here’s a customer:

At the top, we have the person’s email (which I’ve blurred out) and some basic information such as how long they’ve been a customer, the total revenue they’ve contributed, and the first referrer that brought them to us.

And below that, we have the Timeline which shows EVERY action on the site. I’ve hidden many of them to simplify things a bit. Every time an action is completed by this customer, a dot gets added to the Timeline on the day it was completed. The bigger the dot, the more times the user did that action. And when we’re tracking funnels, the Timeline even connects the dots of those steps for us so we can see how they move through the signup process.

This tells us how an individual person is actually using the product. It’s perfect to track different features, sections of your site, products, or anything else you can think of.

What does this Timeline tell us?

Well, it looks like they found us through a marketing campaign, signed up on the same day, created a survey, and then upgraded their account when they received the first survey response a few days later.

There’s also a bunch of activity over the next few days… but then it tapers off pretty quickly. Within a few weeks, this customer has stopped using the product altogether. That’s definitely not good.

So this customer originally found value in the product but didn’t become a heavy user. I have two options at this point:

  1. Get in touch with this user (I have the email) and try to understand how the product could have been more helpful.
  2. Start looking through Timelines of other people to see if this behavior comes up frequently.

If I find that a number of my customers are using my product a lot right after they sign up and interest quickly wanes, I then want to brainstorm options to increase engagement. Here are a few options I might consider:

  • Maybe people are only creating a single survey to answer a single question they have on their site. I could write some emails that show all the ways that people can use surveys on their site.
  • Maybe they expect a certain feature or benefit that I’m not providing. By reaching out to my customers, I can figure out what that is and build it.
  • What if the survey responses aren’t that valuable? I could put together some guides on the best practices for building effective surveys.

I have a number of possible explanations; all I need to do it start talking to my customers, get feedback, and test my ideas. Within a few months, it’s entirely possible to radically improve engagement with the product because I have reports that tell me what’s going on.

When data on individual people is at your fingertips, you’ll find all sort of amazing nuggets on how your customers behave and what you can do to build a better business.

Where to Find Data Like This

What’s that? You don’t believe this analytics fairy tale exists? Nonsense, it’s real and just as awesome as it sounds.

In fact, we built KISSmetrics from the ground up to give you data on your actual customers (not pageviews). All the reports I’ve already showed came right from a real KISSmetrics account.

Funnel Reports

As we already covered, funnel reports are awesome. In fact, they’re more than awesome, they’re awesome-licious.

But Lars! The analytics product I’m using already gives me funnels. What’s the big deal about these?

First, most funnels only give you aggregate data on how people move through the funnel. But we know aggregate data hides really important insights on your business. The KISSmetrics funnels let you segment your data however you want. And there’s no complicated setup either, simply point and click to break your data down.

Second, the KISSmetrics funnel is the only one I know of that tracks individual people. This means you can build funnels that track the real steps of your business funnel instead of you having to design your business around what’s trackable. Putting add to cart buttons, twitter shares, or the use of a specific feature into a funnel is very easy to do. And it doesn’t matter if people do all sorts of other tasks between those steps.

Google Analytics will only build funnels with consecutive pageviews which means you’ll have to mangle it pretty hard to get a funnel with steps like these:

Once you’re collecting data with KISSmetrics, you can build a funnel like this in under 5 minutes.

Cohort Reports

Here’s a report we haven’t already talked about, the elusive Cohort Report. So what in tarnation is a cohort anyway?

A cohort is a specific group of customers. The most common way to break up your customers into different cohorts is by date. In other words, the customers you acquire this month are in a different cohort than the customers your acquired last month.

When your break up your customers into cohorts and analyze behavior, you can quickly see how behavior changes over time.

Let’s say that your customer base has been growing steadily over the last few months. But it’s not because you’re acquiring brand new customers, it’s because the batch of prospects you found in February are still converting into customers. A cohort report will tell you this. But if you just look at monthly conversion data and customer counts, you’ll totally miss it and won’t realize the importance of what you did in February.

Here’s what the report looks like:

People Reports

We’ve already covered the people reports and how useful they are. But go ahead and have a look at this:

This is a list of every bit of data you’ll have on your actual customers. Every search engine visit, every action, and every way they’ve connected with your business.

You’ll also have a complete list of the search terms they’ve used to find you, every referring source of traffic, each split test they’ve been a part of, and any other piece of data you’d like to set up. All of it.

What could you do with this data? Think of how many questions you could quickly answer that will help you take your business to the next level.

Want to Get Started?

Try KISSmetrics out for yourself with our free-trial, these reports will tell you exactly how customers interact with your business.

Bottom Line

Always be careful with aggregate numbers. They do a great job at giving you a quick overview of how your business is doing but they hide all sorts of insightful nuggets on how move things forward.

Whenever you see an aggregate number, try to break that number down so you can see what’s really going on. Start with segmentation (breaking the number down into different groups). Then dive even deeper and start looking at individual people.

So get to it and break that data down!

If you’re a KISSmetrics user, how have you used you data to find interesting trends and behavior on your customers? Tell us in the comments!

About the Author: Lars Lofgren is the KISSmetrics Marketing Analyst and has his Google Analytics Individual Qualification (he’s certified). Learn how to grow your business at his marketing blog or follow him on Twitter @larslofgren.

  1. Very interesting in depth post Lars.

  2. I’m a regular reader of your posts and have to say this is the most compelling article I’ve read yet in terms of getting me to switch from Google Analytics to Kiss Metrics. The more of this kind of stuff you can expose, the better!

    • Glad you liked it Albie! Definitely let me know if you have any questions on analytics or KISSmetrics and I’ll see if I can work them into a future post :)

  3. Hey,
    Excellent post. This is true that real people can increase your sales and improve your business. Most of the business must have target visitors so that they can convert. I would suggest that targeting right keyword can bring a lot of results.

    Thank you

  4. I think other web-site proprietors shluod take this site as an model, very clean and fantastic user style and design, as well as the content. You’re an expert in this topic!


Please use your real name and a corresponding social media profile when commenting. Otherwise, your comment may be deleted.

← Previous ArticleNext Article →