What’s happening to your business, product, or game right now? As the web gets faster, we want analytics and data just as fast. Facebook and Twitter have already built their own systems to measure their petabytes of real-time data. But there’s nothing stopping the rest of us from using real-time analytics, too.
So how do you actually use real-time data?
Let’s keep in mind that actionable metrics help us make better business decisions. We should take the same approach to real-time metrics: actionable > vanity. While there are few use-cases where you can take action on real-time data immediately, there are benefits to having access to data within minutes of an event happening. I’m going to show you four ways to take advantage of real-time analytics.
First things first. How do you know the data you’re tracking is correct?
It’d be a shame to spend time configuring events, segments, and metrics only to find out later that nothing triggered correctly. Real-time analytics prevents that from happening.
The KISSmetrics Live feature shows me real-time data so that I can see events triggering correctly as I move through the site.
When I set up analytics for Crazy Egg, I needed to see if my events and properties were triggering properly. KISSmetrics has a Live feature that helps me do exactly that kind of Q/A. Once I setup my event triggers, I let the KISSmetrics Live feature run in the background as I clicked through my site. Thanks to the real-time data I was able to see if I triggered events and properties as I moved through the different sections of the site. I even decided to do a search through Google just to see if data was being passed in correctly. So far so good!
Real-time data can help your initial analytics setup and implementation be a breeze. Without it, it can be stressful to find out that the report you’re supposed to deliver tomorrow doesn’t even have the data you wanted or needed.
2. Monitoring metrics/campaigns/behavior
Now that real-time data has helped you successfully debug your analytics, you can start having fun with it to monitor usage, metrics, campaigns, and more.
KISSmetrics Live shows Ad campaigns driving traffic to a site and bringing in sign-ups.
Want to see how effective that marketing campaign is that you just launched a minute ago? Now, you can.
KISSmetrics Live shows real-time data as a user arrives from Klout and goes through the Facebook Connect process.
Real-time data not only helps you measure what’s happening on your own site or product, but also measures how your partners or affiliates are driving in traffic as soon as you make that connection. If you have integrations with Twitter, Facebook, Klout, or other sites, you can see if those features are even being used (and of course, test if you integrated correctly).
In a mobile app or game you’ll be able to see how users behave when they first launch the app. If you’ve just launched a new feature, real-time data is especially useful for seeing if your users are adopting it from the get-go. It does not, however, let you see if that adoption led to a meaningful result such as increasing your bottom line or retention rate. More on that later.
3. A/B Testing
High volume products such as social games or apps can optimize their products within minutes with real-time data.
Zynga can test the Cityville “Add Coins & Cash” button for thousands of users within minutes.
Let’s say you wanted to test a couple different versions of a new “Purchase” button to see which one received more in-app purchases. A developer can setup a quick split test, let it run for a few minutes (or hours, depending on your amount of traffic), and stop it. Within a few minutes, the data he/she would have would be significant enough to make a decision. And similar tests could be repeated over within only a couple hours to find out the most optimal flow, button, or layout for the product.
The ability to do this is valuable for teams that iterate features multiple times in a day because data will be available immediately as iterations are rolled out. If you need to test fast (or maybe settle a dispute), the data will be able to let you determine a clear winner.
Amazon’s can do A/B testing for the new Kindle by switching the navigation of the Kindle to the top.
What about e-commerce sites?
Amazon.com is a great example of how real-time data can be great for split testing. Amazon’s familiar layout is due to years and years of optimizations and tests. They have a large amount of baseline data that tells them what works for them. They probably won’t make any major changes.
But since Amazon gets millions of visits a day, they could run a simple test to see if switching the Kindle sidebar option with the Unlimited Instant Videos leads to more traffic towards the Kindle. It makes sense since they launched new versions of the Kindle and announced the Kindle Fire. This test could run for just a few minutes to get enough data.
Remember, all these tests are meant to push you into action so you can make a decision. The benefit is that you can see this data as it flies in and get things done faster. Who doesn’t like getting things done faster?
4. Content based on user preferences
What do Amazon and The New York Times have in common? Real-time content publishing. When you use real-time analytics, you’ll be able to tap into user preferences as people are on your site or using your product. By knowing what users like while they are on your site, you’ll be able to take action by showing content that is relevant. It’s a smart move. If you give your audience more of what they like, that raises the quality of content on your site. And that increased quality gives a better experience overall. Let’s take a look at how this works for these companies.
Amazon Recommendations change after each new product you view so that they can upsell customers throughout the session.
Ever go back to the Amazon home page after you’ve done some browsing? Amazon has at least five sections of recommended items based on your browsing history: New for You, More Items to Consider, Related to Items You’ve Viewed, Inspired by your Browsing History, Additional Items to Explore. How do they do this? Amazon has spent almost a decade perfecting their algorithms to process data in real-time and give recommendations during a user session. When Amazon set out to solve this problem, they wanted to target these main challenges:
- A large retailer might have huge amounts of data, tens of millions of customers and millions of distinct catalog items.
- Many applications require the results set to be returned in real-time, in no more than half a second, while still producing high-quality recommendations.
- Customer data is volatile: Each interaction provides valuable customer data, and the algorithm must respond immediately to new information.
What we know now, as Amazon Recommendations, has been the result of their developments. And I would say it’s a huge success.
So how does Amazon benefit from using real-time data? Targeted marketing. Real-time recommendations create a personal shopping experience for each and every customer. With more insight into their customers on an individual level, Amazon is able to effectively upsell and cross-sell products at every interaction point. If there’s one thing we should learn from Amazon’s success, it’s that we should use real-time analytics to make the customer experience better.
The New York Times
The New York Times homepage featured articles change multiple times a day.
How does the New York Times decide which story to feature more prominently? The New York Times pays attention to their reader behavior using real-time analytics so they know what’s being read at any time. This helps them decide which position a story is placed and for how long it’s placed there. I keep returning to The New York Times not only for the quality of the work, but the relevancy.
Services like Newsbeat and Visual Revenue help major media sites handle content tracking for every article as well as front-page automation. And this is useful because media sites are interested in growing traffic, increasing engagement, and gaining audience loyalty. Real-time data allows publishers to see what is trending throughout the day so that they can deliver the right content at the right time. This also lets them know which articles may be spreading more virally as opposed to being discovered on the site.
The longer a reader stays engaged on the site, the more informed a publisher can be about the user’s reading preferences. And the longer a reader stays engaged on the site, the more likely they are to become a return visitor. With real-time data on a reader’s preferences, the publisher can recommend articles that might make them stay on the site longer and engage further.
At the end of an article, The New York Times has a fly-in modal to give readers another article to read and increase engagement.
Is Real-time Worth the Hype?
It’s hard to think about common situations where real-time data accuracy (<1 second) would be useful. An exception would be a trading platform that depends on bidding at high frequencies. That being said, access to minute-by-minute data can be very useful to businesses, as I have shown you. Timely access to data can save time, money, or reputation (if your service goes down).
When it comes to the decision-making process, it makes more sense to look at historical data. How often can you make a split-decision based on data you just saw? You’ll usually want to see the relationship of multiple events as they relate to your revenue, retention, or other metrics that are important to your business.
That’s why earlier I called it real-time “monitoring” as opposed to real-time “analysis”. Real-time data is great. It allows me to find out what’s happening right now and investigate deeper with more appropriate tools such as a funnel or cohort report. It’s a welcome tool in anyone’s data arsenal.
Remember, the most important thing to ask yourself whenever you measure data, real-time or not, is “What is the purpose of this data?” If you’re using real-time data, measure what matters – not what looks good for a press release.
About the Author: Chuck Liu is the Product Marketer at KISSmetrics and cooking enthusiast. When he’s not being a nerd writing about data, he’s in the kitchen whipping up something delicious. He also loves corgis. Follow him on Twitter @chuckjliu.