Do you use Google Analytics? What are your pain points? Things you think might be more achievable if you only knew how?
I use Google Analytics but don't really have a ton of experience with it.
I wish so badly that the data was in real-time all the time. I get that ETL is a big deal for larger sites/companies. For small startups, it should be trivial to have a real-time analytics solution.
I can't really complain about the complexity of it because I haven't put in any time to learn it really. They offer free courses, I just haven't done them.
The gtag script is really easy to use. It's nice that it only has a single function call to send events. However, the documentation about what the fields are for those events is confusing. I still don't understand how action, label, and category are all that different. It's also not straightforward to view those events and filter them because most of the default charts only take two dimensions.
In my day-job we use Splunk to monitor systems. I like that model of every log is an event. They provide a proprietary query language on top of it. The flexibility is very powerful. I feel like Google Analytics limits this to try to be simple for people who don't write code.
Have you ever worked with ELK stack? To me, it seems like a great replacement, but I only used Splunk at one job and only infrequently.
Never heard of it, but just checked it out. It seems like an open-source replacement for Splunk. The query capability of Splunk is pretty powerful, but I imagine that this ELK stack could easily match it. I'll have to try it out on my own!
I work at a big company so the splunk decision isn't mine to make. They had already paid for it and use it company wide long before I was even there.
Same. I was at a unicorn and they paid for all kinds of stuff I never could on my own.
First three paragraphs I get and agree then I drop off, events and the complexity to go with, I just haven't dug deep into.
I use Google Analytics to estimate how many people are coming to a site a day or month, and make sure that trajectory is uphill. I'll dig deeper sometimes to see which landing pages are bringing it in or which queries.
Working as a single dev for a small agency, If I'm lucky I'll have set up conversion tracking (I do this for my own good not as a paid offering so I'm not allotted proper time). Then if I'm real lucky later I can tell which keywords or landing pages are converting the best.
That's the extent in which I use analytics, I do feel like there is much more I could be getting out of it but I just don't know and have to spend time going back and seeing why conversions aren't working, would be nice to know for sure.
Interesting. If you went the enterprise route and paid for GA360, it would be near-realtime (within 15 minutes). You'd then be able to set up a stream of hit-level data into BigQuery which would give you a much more Splunk-like environment.
I think if you added the tag to your site using Google Tag Manager (so add the tag container), configuring the GA tag would make more sense as it's more wizard-like within GTM - so fields, etc are selected rather than being things you need to type. GTM does come with it's own learning curve, though.
Oh, I see. I, as an IndieHacker, am nowhere near being able to afford GA360. I only use Splunk at my day job which is a large company.
I don't quite follow about using Google Tag Manager. I recently just set analytics up on my site about 1 month ago and it was my first time ever doing that. I didn't explicitly set up Google Tag Manager, but the auto-generated code snippet it gave me for my head tag was the script that loads gtag.
I don't really like wizard like things very much as I prefer to code and have all of the control. I just wish the explanation of how to use category, label, and event were more clear from the GA docs.
Only one conversion per session (per objective). Limited custom metrics and dimensions. Slow interface and bad UX IMO
You can have several conversions per session/objective - without knowing your use case I'd guess that you'd be best reporting on this using goal value (or cart value if it's an e-commerce site).
Custom metric/dimension limits are in place to persuade companies to go with GA360, which has far more slots available.
I feel your pain on the interface speed and UX!
I am using statcounter, because I am to lazy to set up google analytics. Perhaps in future I will use it, but my main goal is to communicate with a laser targeted audience and have not so much crap traffic, but a targeted one...
It can't be that much more effort to plug in GA over statcounter - assuming you're not configuring goals or customising settings like session duration etc.
No, I have my own platform.
One you've rolled yourself? Well done, I've always been keen to have a go at this. How did it go? Harder/easier than expected? Do you feel you've been able to code all the features you need?
If you want all data in one dashboard, you must go this route. All the other products are canned in such they do what the majority of folks want, not more. The costs on an ongoing basis are very tiny. The coding is not that big of a deal. You can customize what you need exactly. And we noticed all services are a bit inaccurate. With our own code you can tweek it.