Many of us use Google Analytics as our day-to-day analysis and reporting tool, it’s provided enterprise level analytics to everyone, and turned a legion of website owners into quasi-statisticians.
However, it’s not without its flaws and weaknesses. As I’ve been a Good Boy this year, here are the ten things I’d love to have from Google Analytics for Christmas.
Now, before starting my list, a couple of qualifiers. Firstly, yep, I know that Google Analytics is a free tool. However considering how much money I throw Google’s way in PPC fees, I think I’m allowed to suggest some ways they can improve it.
Secondly, this gets pretty heavy. I’m a statistical analysis nerd and I make no apologies for it. However, if you’re at the stage where you’re starting to use some of the things I mention here, then you’ll reach the same frustrations pretty quickly!
Make metric & dimension combinations easier
Google Analytics Custom Reporting allows you to knock up reports pretty quickly, dragging metrics and dimensions into a report to get exactly what you want.
But, if you’ve ever created a Custom Report, or use the Google Analytics API, then you’ll know very well the frustrating, seemingly random and illogical restrictions on which metrics and dimensions you can combine.
Often I’ve known exactly the report I want to write, only to be left banging my head against my desk because the desired metric is suddenly unavailable, or you get the blasted “Illegal Combination” message from the API.
Sometimes it gets even weirder, and you’ll be locked out of reporting on e-commerce metrics, but goal metrics are just dandy. I’ve been using this system for a long time, and I still don’t understand the logic. Maybe I’m just dense. It sure is frustrating.
For example, writing a report that shows Site Search Activity by Landing Page, you’re locked out of e-commerce, goal & event metrics. So you have no idea how successful the each search was, which would give you indicators of how to curate the landing page.
The only way is to create a goal conversion by landing page report, then set up some segments for the search terms. Laborious and unnecessary.
There are even separate guides for the GA interface and the API. Here, see if you can figure it out.
“bangs head on desk”
There’s an awful lot in beta at the moment. Google does love its betas. Intelligence, In Page Analytics, Adwords are all in beta. Sometimes they work well, sometimes they don’t work at all (I’m looking at you, In Page Analytics) and replace perfectly good functionality that you now have to do without (looking at you again).
The Google Analytics interface has become terribly slow lately. Maybe it’s just me, but applying a new segment or data range means sitting there with the spinning spinny thing of death (its official name) going around for a good 30 seconds. But it’s not so bad, you can use the API! Oh, wait.
A robust API
Google Analytics provides a shiny API that allows you to draw down data for your own analysis. I do most of my analytics work using the API nowadays, I can product exec summaries and KPI dashboards super-quick. Except I can’t because the API is flakey as all hell.
Whether it’s because there’s no servers to execute my request, timeouts and the threat of going over quota, it makes using the API as a reliable Dashboard or dynamic presentation tool impossible.
It’s even worse in Google Docs, where if anything, service should be tiptop. The Google Analytics API functions should really come as standard in Google Spreadsheet. I’m amazed that Google hasn’t integrated its business apps as tightly as it could.
Less sampled data
So instead of pulling down lots of data, I partition my request to avoid timeouts, so say 10-15 smaller requests, filtered on pre-defined segment, with maybe three metrics per request. But when I then try to pull it down for more than two weeks…DISASTER.
The data returned is sampled, and therefore pretty useless to me. Sampled data is fine for looking at trends, not so fine when trying to do a 1, 4 & 13 week comparative analysis.
More charting options
So, if you can’t use the API as much as you’d like, then you’re back to the GA interface to produce your whizzy reports. However the reporting options are pretty limited, and you’re screwed on what would be some quite useful reports.
For example, what I’d like to see, based on a date range, is a histogram of transaction values, which would show me skew, outliers & deviation. Yum. Where i work, for security & privacy reasons, I have zero access to any CRM database, so GA is my first stop of this sort of data. Why can’t it be better, dammit!
Manageable advanced segments
Advanced segments are great. Little subsets of data that you can report on. Once you start using them, you’ll quickly keep adding more and more.
Until you reach 100 of them. That’s all you get.
But, that isn’t the worst part. Long before you reach 100, you’ll realise how completely unmanageable Advanced Segments are.
You can’t sort them, you can’t group them, you certainly can’t have subfolders of segments (ooooh, advanced segment folders with parent inheritance, wouldn’t that be cool!). You can search, but I’m a bugger for not having a naming convention for these things.
More major contributors in Intelligence
Google Analytics Intelligence is great. From what used to be the world’s tardiest alerting system, by now reporting on the major contributors on alerts, GA has built a sneaky multivariate analysis engine under our very noses.
How about a few more variables now?
I would love to see contributors on Conversion Rate & ATV. Revenue’s nice but you’ve then got to work backwards from it. Yes I can work these out (and indeed I do using a Metric Investigation Framework Google Doc that draws that alerts data down from the API, when it works), but to have it at a glance would be great.
Calculations in reports
Custom reports sounds great in theory. You can build up a report, set a distribution list and schedule, and have GA send out regular Exec Reports to your top brass. Sadly, Reporting system isn’t Exec ready.
In custom reports you don’t have access to calculated fields, which means you can’t report on a lot of important statistics, like Conversion Rate, but instead have to download, manipulate (Google provide a handy reference guide for its calculations) and distribute which is a real shame.
No more URL hacks!
Right, leave the best to last Matty. Occasionally, you’ll find yourself limited by the Google Analytics user interface, you’ve come across some silly restriction or random behaviour of the interface.
The two most common ones are:
Downloading more than 500 rows of data
Removing All Visits from Selected Segments (GAH!)
But, don’t worry, Google has actual authorised “workarounds” that involve diddling with the URL. Here’s how to get more rows.
Right, the All Visits “feature”. This is probably the single most annoying thing about Google Analytics. I’ve complained about it until I’m blue in the face, but to no avail. You’re selecting some advanced segments to report on. You untick All Visits, and tick two others. Look! All Visits becomes ticked, and you can’t untick it!
Even if the segments aren’t exhaustive, so Segment A and Segment B don’t cover the entire data set, you will always have All Visits in your report. Fine and dandy and you can ignore it. Unless you want to see a graph, where All Visits makes the other plots pretty useless.
But, hey, you can hack the URL. Look for a bit that says “seg0=-1 and delete it. Huzzah, it’s gone!
But wouldn’t it be better if you didn’t have to?
So there we go, my Christmas Wishlist for Google Analytics. Hopefully Santa Kaushik is listening. Whilst we’ve got his ear, anything you’d like to add?