I recently discovered that users of Google Analytics Free don’t always fully understand what they’re seeing when they log into their dashboards. They think they’re looking at their actual, honest to goodness data. So let’s clear some things up:
1. Your data is fake.
“What are you talking about! It’s not fake, there are graphs!” Your data is sampled. In some situations it’s still statistically significant. In fact, in most cases it is, so don’t panic yet.
2. Want access to your real unsampled data? Fork over $150,000.
For major corporations it’s pocket change. For the average person with a small site, your sampled data is still statistically significant, so it’s no big deal. If you’re a start up company who is still poor but is experiencing rapid growth, you’re out of luck. Break out your wallet.
3. If your company is expecting/experiencing rapid growth & has set up a bunch of custom events, of the four leading analytics companies, Google Analytics Premium is by FAR the cheapest.
If you hit 5 billion events per month, Google Analytics Premium bumps up to a flat rate of $220k. Adobe Analytics jumps up around $7.4 MILLION dollars per year. That is not a typo, it’s a legit quote. If you’re scaling crazy fast and need quality data, Google Analytics Premium is your friend from a financial perspective.
4. Unsampled, pure data is freaking expensive.
It just is. If you really need it, surrounded by a UI that isn’t terrible, be prepared to shell out some serious cash.
5. I previously mentioned that sampled data is statistically significant for most companies, so who cares about whether data is sampled or not?
CMS companies in particular need to care. Let’s say a company has global analytics tracking. Their 4 largest clients may hit 3 million events, 2.5, 2, and 1 million events per month. They may also have a handful of 100k — 250k events-per-month clients and so on into tiny client land.
Google samples the daylights out of your data. The higher the number of monthly events, the worse your sampling rate becomes.
From a global tracking perspective looking at high level metrics, sampling is no big deal. However, if a 100k events-per-month client calls and requests a report about mobile device traffic to their primary domain, you’re statistically SOL. You can pull the report, but the drill down data will be garbage due to the crazy sampling rate.
The worst sampling rate I’ve come across in Google Analytics Free was .02%, but you get what you pay for.
The moral of this story:
For small sites and small businesses Google Analytics Free is a viable option. Your data is “fake”, but still statistically significant.
If you’re a rapidly growing business with a pile of domains being tracked under one account, eat the cost and upgrade to Google Analytics Premium. From a scalability perspective it’s the cheapest option, and you can pull unsampled reports to obtain accurate data all day long if you want to.