I use Fathom[1] for simple, privacy-conscious analytics on my blog. It is interesting to know which pages people are visiting and where they're visiting from. It's rewarding to see month-over-month increases in traffic, too--even if I don't get a lot of traffic. It's motivating to see that people are spending their own time to read what I've written.
Different people will use analytics for different purposes, and different metrics will matter to different people. For me, seeing people read multiple posts (when the number of page views is greater than the number of visitors) is the most rewarding metric. It's one thing to click through from Google. It's another to click through one article and want to read more.
I saw Fathom a while ago and it looked very useful, but the fact that it only supported a single site per instance was a non-starter for me. Have they changed that?
I guess it all depends on your reasons for starting a blog. The author of this piece is a marketer, so is very interested in gathering metrics to see how well they are presenting themselves.
Personally, I think slavishly following metrics leads inevitably towards low-content click-bait articles, something that this article seems to be actually advocating for. The silliest things will get a lot of hits while more considered content will get ignored.
Additionally, although there are worse things than Google Analytics, you are imposing it on your users for your own gratification. Perhaps your users don't want Google to know every page they visit.
Here is my advice for starting a blog: start a blog.
Install GA if you want to look a pretty graphs but don't worry too much about the numbers unless you are selling something. And if you are selling something, worry about sales, not hits.
Make sure you choose a goal before choosing the metrics you care about! The metrics OP lists look like they have a goal of driving traffic and engagement which is good for some blogs. But for most of our (hn-crown) blogs the goals look like "sharing info with others", "solidifying me as an expert in this space", or "storing info so I can get back to it later".
I think that analytics on small websites does not make sense anymore! A lot of people have adblockers these days and those often block Google Analytics for privacy reasons and also just for speed/less work.
Also with bots that execute javascript with Electron, it is nearly impossible to have correct metrics coming from anonymous users.
A better metric would be to measure newsletter signups, third-party client-side analytics is dead.
Long live awstats, the perl script that you put in crontab to analyze your apache logs. I have like 15 years of data from this and it's still chugging along. Not overly beautiful but it gives me the first order metrics. Traffic, referrers. No google needed.
Edit: I guess this doesn't help with the anonymous bots issue.
This article was a bit of an advertorial for the blogging platform.
Collecting analytics is the easy bit. Even if you have not done it before you can get the job done with the Google Analytics approach with an hour of fiddling about, with no prior knowledge and from a standing start. If you want to do something server side or implement whatever Piwik is called these days you just have to follow the instructions. It is not a hard problem.
Content is king but I am questioning whether this blogging platform is deserved of fine content, regardless of how the clicks are measured.
The problem I see is that the platform lacks decent HTML vocabulary. It is not difficult to structure an HTML document and a platform should look after that aspect so people don't have to be an expert and can just focus on the content. But here I see the platform is selling people short. The basic HTML is not up to scratch, which is a fair thing to say when accessibility is compromised.
The page does not use a header, a main or nav elements. Yet it has what visually appears to be a header, main content and some navigation 'landmarks'. These should be surfaced with the correct HTML5 for the job rather than obfuscated in div groupings.
The images are not very accessible, they should be figures as that is what they are. The comments should also be 'articles'.
Medium may not be perfect but at least the HTML does get the basics right.
I don't see the point of having these websites that use fancy React thingamies and Bootstrap same-as-every-thing-else buttons but don't get the most elementary HTML right. Why can't people keep things simple?
Having explored the hash-whatever thing more it actually is some of the poorest HTML out there. This doesn't matter to most people but, as mentioned, you trust a platform to get this bit right for you, as is the case with Medium.
The article claims that browsers drop referrer (well, "referer" :) header when you open a link in a new tab (look for those direct hits mention).
The last time I checked, that was not the case: can anyone confirm if this is true (I'll check myself later, but curious if someone knows off the top of their head).
I think most browsers don't include referrer information when clicking a link on a HTTPS site which leads to a HTTP site, but other than that I believe the information is kept by default.
If you're on WordPress and value your users privacy, do not use Google Analytics.
Sure, webmasters get to take advantage of big brothers' trove of fingerprinted browsers and fancy metrics, but you're also diming out your users to Google and feeding a beast that's already obese.
Google benefits most when you put Google Analytics on your site. Now they can track your users. Webmasters getting analytics out of the deal is the only way Google can sign you up to spy on your traffic.
That's why I use the 100% self-hosted WPStatistics plugin on my WordPress.org blog. And it's considerably faster than Google analytics too, because you don't have to make your clients do an extra DNS query and download external tracking scripts from 3rd party (Google) servers.
Web analyst here, not Google employee although I put bread on the table with GA.
I am reasonably confident that Google does not dip into GA data for its own purposes, unless you go out of your way to enable Advertising Features. Google disclaims ownership and controller-ship of the data collected by GA. This is a contract to its users, as well as the cornerstone of Google's plan to not get fined into the mantle of the earth via GDPR.
There are valid reasons to avoid Google Analytics, including privacy ones. But the idea that Google uses GA data for profiling is overblown (again, aside from Advertising Features, which establishes a connection between GA data and advertising data). It has plenty of other sources for personal browsing data, and those other sources are more reliable.
And since this always comes up: Google gives GA away for free because it's a strategic compliment to online advertising. If people can see how much ROI they get from AdWords, they spend more on AdWords. It's that simple. Google does not need to dip into the GA data in order to see a massive return.
I hear you loud and clear. I'm still not willing to risk it though. I had people calling me crazy when I started complaining-out-loud that Facebook was asking for my phone number "for account recovery reasons." Of course the word of Facebook is sacred!
How much faster is "considerably faster" in this case? GA is one of the most universal and most responsive client loaded dependency. Is your site so lightweight that this makes a difference?
You are giving google more information about those users by letting them know that they visited your site specifically and how engaged with your site they were.
We've been using goaccess [1], we've been very happy with it and most important, we keep things private. Installing Google Analytics is def a NO for us.
I've been planning to switch to goaccess (from piwik/matomo) for my personal sites. Do you have any tips? Generally I just want to know trends in unique pageviews/visitors; whether any page in particular is seeing a lot of referred traffic and from where, and what paths people are taking through the site once they arrive. I like the purity of going solely with log analysis, but I'm a bit concerned I won't get quite as much visibility.
I'd say start with the basic command of goaccess and then you can add extra options to the command line as needed. We have ours on a big display at our office running real time with only the most relevant panels that fit our needs.
Overall the stats are pretty accurate, we did some tedious/manual analysis of our weekly access log and the counts displayed by goaccess are on a par with our data results.
"If you're just getting started with blogging, make sure to set up Google Analytics"
No. Please don't add Google spy code to your site.
If these sorts of analytics are actually important to you, consider using an analytic package (or, better, write your own or parse the information out of your server logs) that is entirely hosted on your site and doesn't report to a third party.
Recommeding Google Analytics isn't really great. Google already has enough data and you shouldn't sell your visitors data to Google just because they offer "ananlytics for free". Thanks to UBlock Origin Google Analytics get's blocked by my Firefox.
Because I wasn't happy with any already existing analytics, also Fathom because they use cookies and I wanted to avoid this, I build my own KISS statistics tool called KISSS or kis3 using Go. It has minimal functionality, is easy to run everywhere and requires minimal resources. See https://kis3.dev
analytics were valuable / helpful when they included keywords and search phrases used with google - ever since google starting stripping those away, I've spent little time looking at analytics unless some kind of attack issue or something.
I wish google had a privacy option to let users decide if their search phrases should be sent / captured by the web sites people visit. One could dream that they are auto stripped if opening in incognito mode and sent in regular mode unless another setting was changed by the user in chrome or via google account or such.
There are a lot of assumptions baked into this post that will not universally apply. I look a bit at things like page views, but I'm much more interested in other things. Looking at the numbers can help me tease out a few things, but those numbers are generally secondary to how I parse things.
Different people will use analytics for different purposes, and different metrics will matter to different people. For me, seeing people read multiple posts (when the number of page views is greater than the number of visitors) is the most rewarding metric. It's one thing to click through from Google. It's another to click through one article and want to read more.
[1] https://github.com/usefathom/fathom