Analytics for Community Growth - The Good, Bad and the Ugly
Jatra's latest upgrade has a lot of goodies packed in. I'll write about all the SEO improvements via a separate post. I observed something that's bothering me for a while - and I wanted to discuss it with fellow Jatra members.
In the earlier version of Jatra, we opted for Google Analytics (Version 4) to track community growth. GA gives you a lot of data - 90% of which is never used or tracked. Serious businesses have moved to product analytics with services like Mixpanel or PostHog.
I made an observation with GA4 and it's not what I expected.
The Trigger
When we add Google Analytics tracking code to our pages - the page-speed score drops from 90s to the 80s or sometimes 70s. Remove it - and performance goes back up to ~95.
Note: Jatra is already fast. We opted for Server Side Rendering (SSR) instead of relying on JS to render pages. The PageSpeed score in 80s it not bad at all. Even the score in 70s is acceptable
The point is - Google Analytics adds an extra DNS ping and is uncompressed. We tried optimizing Google Tag Manager (required for Analytics) - but we couldn't cross the 90s.
The Good
Google Analytics gives you lot of important data - if you are really data-minded. You can see which geo-locations are visiting the site, retention, bounce rates, top content et al.
If you are a community builder - it means reduced guesswork.
Analytics also helps defend the ROI. You can present your numbers to clients / leaders for credibility.
The Bad
While analytics is good - introducing it on your community adds weight. Third party scripts, DNS lookups and background processing adds up.
Even when loaded asynchronously, they affect performance. If speed influences rankings and rankings influence growth- the cost is trivial.
You are also sending your data to an external platform and rely on their fancy dashboards and definitions.
The Ugly
Most community managers track vanity metrics. I get that. Growing count of pageviews makes community leaders happy. Bounce rate does not reflect whether someone found the discussion valuable.
Should We Build Our Own Analytics?
That made me wonder - if we should build our own analytics. We already are able to track data accurately; but information like geo-location, traffic sources will require introducing additional JS. We can bundle it with the app.js that your community downloads to eliminate extra DNS ping.
However, it also carries opportunity cost. It requires event architecture, storage, processing, visualization layer and privacy safeguards.
My questions to the community:
Do community builders really need generic web analytics or community-focused metrics?
What data do you track in your community and why?
Building analytics is a significant investment; and I'm wondering if I'm thinking in the right direction.