Analytics Market Chart

Real-Time Analysis and Google Analytics

Why Real-Time?

Here are some of the most legitimate reasons for real-time tracking that we have heard:

  • Need to know if ads are underperforming in real-time so that they can be modified or discontinued.
  • Need to know if a page is failing so that it can be remedied.
  • Want to track individual users as they interact with my site.
  • Need to know if a page is popular so that a link to it can be posted somewhere.
  • Want to proactively engage visitors on the site.

There are probably a few other common reasons, but this list probably summarizes the bulk of them pretty well.

Who Does Real-Time Reporting?

Strictly speaking, no web analytics software to my knowledge does true "real-time" reports. All of them lag behind by at least a few minutes. As of this writing, Google Analytics lags behind less than an hour on average. Officially, it can take up to 24 hours. Sometimes reports show up in two or three hours.

There are programs that offer true real-time reports. Tealeaf is an example. It sits on a port of the webserver and processes vast amounts of data for your site in real-time. It can even send alerts if certain thresholds are met or a certain combination of events takes place. But Tealeaf is not web analysis. Nor is most other real-time reporting software.

Web Analysis vs. IT Reports

It's important to differentiate between web analytics and IT reports. Web analysis consists of spotting trends and correlations and answering big picture questions. It is used by website owners and marketing professionals to make decisions about which pages to optimize and which marketing campaigns are underperforming. Web analysis is all about creating a list of things to work on that should increase conversion and visitor satisfaction. It is focused on people and trends.

Data for web analysis is gathered specific to that purpose, not for error reporting. The methodologies are intentionally and necessarily different.

IT reports, on the other hand, focus on hits, server load and file errors. If all you want is a hit counter, IT reports are your solution. They are used by developers to make quick decisions about scripts that need to be fixed. They are focused on server capabilities and coding, but not on increasing conversions, per se. If a coupon code isn't working, you need to know now. If a page is hanging, you need to fix it immediately. If the site goes down, you need an alert in the middle of the night so you can fix it.

In other words, if your IT reports aren't real-time, you may be missing some critical information.

But web analysis is based on trends and surveys. It's like asking for a Gallup poll that shows Americans' attitudes toward cheeseburgers at every moment of the day. Any minute-by-minute fluctuations will be statistically irrelevant, and decisions based on them would be flawed.

How Analysis is Used

It's important to recognize that all analysis is post-analysis. Error reporting can and should be done in real-time, but reading into the data to find bigger pictures requires thinking and time. It always happens after the event occurred.

Assume for a moment, though, that somebody actually could relevantly analyze data as the data crossed their screen. If real-time analysis were critical, wouldn't you need to have that person staring at the screen 24 hours a day? If not, why is it only important to have real-time analysis from 8-5 every weekday, excluding holidays?

Taking Immediate Action

What about taking immediate action on marketing or site optimization? Won't an underperforming campaign cost money if it's not addressed right away?

It might, but how do we evaluate whether a campaign is underperforming? We have to look at trends. Just because a campaign hasn't resulted in a conversion in the last 15 minutes does not mean it's a failure. If it's been several hours without a conversion, that might start to mean something. But, even this decision has to take into account multiple metrics and some deeper analysis. It takes data to come to an informed conclusion. Data takes time to accumulate. If you edit or cancel your campaign every hour based on the previous hour's data, you will be making very rash decisions. You are probably wasting more money with that approach than to leave it alone for a day.

To take another example, what if the visits to one of your pages abruptly spikes? What can you do with this data? It's interesting, to be sure, but is it actionable?

Finally, most organizations are not in a position to make immediate changes to their site or marketing based on the numbers feeding in at this very moment. It takes some time to prudently make changes, test them and then push them to production. If you're working off of real-time data instead of analyzing longer trends, you will be asking your marketing or development team to change what they're doing even before they have implemented the changes you requested.