One Tuesday morning at 9:14 AM, my six-month-old article got 37 views in 20 minutes. DEV.to's dashboard just said "+37 views". No context. No cause...
Some comments have been hidden by the post's author - find out more
For further actions, you may consider blocking this person and/or reporting abuse
But wow, this is a great tool!
As for those articles people scroll through instead of actually reading: I always think that if something is genuinely read by 30–40 people, that’s already a successful article. That’s a full room at a meetup or a workshop, after all!
Exactly! That's the shift this tool gave me. Before, I'd see "139 views" and think "meh, not great". Now I see "maybe 35 actual reads" and think "that's a packed room of people who chose to spend 5 minutes with my words".
The room analogy is perfect. Would I rather speak to 1000 people scrolling their phones, or 35 people genuinely listening? The answer became obvious once I could see the difference.
Thanks for getting it :)
That's not only on dev.to. That's pretty much everywhere online. These days, I finished the book, Smart Brevity And that was its #1 lesson: adapt to how people read by writing shorter, clearer pieces.
Interesting, but my data shows the opposite. My longest articles (>10 min) actually get slightly more engagement than my shortest ones. And my most personal pieces (6-10 min) generate 7.3% engagement vs 2.6% for short technical posts.
Most people skim, yes. But the people who read longer pieces engage much more deeply. I'd rather write for 30-40 real readers than optimize for 1000 skimmers.
Smart brevity works for corporate comms. For conversation-driven writing? Depth beats brevity.
This is really awesome! I built a small static mirror in GitHub if you or anyone else is interested in boosting AI searchability on dev posts. I've found it helps improve traffic from outside sources. I'm curious to put the two together and see if it's increasing readability at all. Thanks!
Thanks for reaching out, Ashley! Your tool for boosting AI searchability sounds like a great addition. Since I’m currently focused on tracking the data Dev.to doesn't provide, I’d be interested in seeing if we can combine our approaches. Adding a lightweight tracking component to your static mirror could be the perfect way to actually measure that boost in readability you mentioned.
Thank you so much for sharing this thing! As a data geek myself, I can't wait to try this, seems like I have weekend plans now.
Need to play around a bit with a RaspberryPi, maybe I'll even build a small dashboard with this. I could imagine that certain companies posting on here would also love this as a web-accessible dashboard and insight tool.
Perfect use case for a Raspberry Pi! SQLite + Python + cron runs great on minimal hardware.
If you build a dashboard, share it — I've kept mine terminal-based on purpose, but I'm curious what others will do with it.
The real value is for individual authors who want to understand their writing, not companies optimizing metrics. But open source means people can take it wherever they want.
hi, thanks for this interesting post and the tool.
The results are quite interesting, indeed.
Thanks! The uncomfortable truths were the most valuable part — turns out observation beats optimization.
Thanks for sharing Pascal. You have the numbers to prove my assumption that a good headline is where 80% of the work lies when writing.
Interesting take! The data actually shows headlines get attention, but content drives engagement. I've had great headlines with shallow engagement (clicks but no reads) and mediocre headlines with deep conversations.
Headline opens the door. Content keeps people in the room. Both matter, different reasons.
Thank you for doing this, I can now cross it off my list. Ever since I did my own version of this, some years ago, I always wanted to bring the stats into the Terminal (where I spend a lot of time). I'll give this a try, but if I don't like it I'll make a "competing product" 😁
Thanks! I'm curious what "your own version" was — the linked article is about app monitoring (Sentry/Prometheus), which is a different beast than DEV.to analytics. Or did you mean a different project?
Either way, build a "competing product" if this doesn't fit your needs — that's exactly how good tools happen.
True, I had a different target at that time but your approach got my curios. Also cool that you take the time to interact with comments. 👍
Good work
Such a fascinating journey into understanding your articles beyond the surface-level stats. The idea of tracking temporal data and seeing the story behind each article’s life is brilliant. It’s not just about views, but about recognizing engagement rhythms, subtle shifts, and long-tail conversations. This tool is a great reminder that real insights come from patterns, not just numbers.
Thanks! The shift from "numbers" to "patterns" changed everything.
The uncomfortable truth: most engagement is shallow. But the small percentage that's real creates the actual value — loyal readers who think about what you wrote and come back.
The tool just made those patterns visible through the noise of totals.
Fantastic tool with an in-depth post to explain how and why it came about - great stuff!
I'm in a lucky situation at the moment that the majority of my posts are about my learning journey, so have extra value to me. If any one else reads them and gets inspired - so much the better! As I progress though, I'm sure there will be a shift in my writing, so this gave me some great insights into what to expect and why.
Brilliant work, Pascal - well done!
Thank you! And I love your approach — writing for your learning journey first means you're already writing for the right reasons.
What's interesting is that when your writing shifts (and it will), this kind of tool helps you see how it shifts, not just that it shifted. You'll see the moment when your "learning journey" posts start attracting different readers, or when a more polished piece gets more views but less real engagement.
It's not about optimizing. It's about staying aware of what you're actually creating, even as it evolves.
Good luck with the journey — and if you end up using the tool, I'd be curious to hear what patterns emerge for you!
Thanks for the lovely reply, Pascal. When the time comes, I will be giving your tool a look and will get back to you!
Looking forward to hearing about it! The learning journey → established voice shift is always interesting to observe.
This is really impressive - nicely done!👏🏼
Thanks! SQLite + curiosity = surprisingly useful insights 🙂
Pascal, this is evergreen. You’re not chasing numbers ...you’re documenting a way of thinking about analytics that will still matter years from now. That’s rare, and it shows.
Thank you, and coming from you — the person whose foundation made all this possible — that means a lot.
You're right that this isn't about numbers. It's about noticing patterns, accepting uncomfortable truths, and understanding how things actually work instead of how we wish they worked.
Your initial script gave me the tools. But the real value was realizing I wanted to observe, not optimize. That shift in mindset is what this article tries to capture.
Thanks for building something worth extending. Open source at its best: someone lays a foundation, someone else builds on it, and the ideas keep evolving.
🙌
Thank you for sharing the repo!! 🙌 🙌
Thanks Ali! And yes — @gnomeman4201 shared the initial code that inspired me, so I'm just paying it forward. That's how good tools evolve: someone builds a foundation, someone else extends it, and the cycle continues. Open source at its best 🔄
Interesting! I am always up for reading an article if it grabs my attention. You definitely piqued my curiosity with this analytics-pro dev tool — I’ll have to give it a try. Thanks for sharing!”
Thanks! Quick note: "analytics-pro" is @gnomeman4201's original tool. Mine extends it with temporal tracking. Both worth checking out — his laid the foundation for everything I built on top.
ok! I see. Cool!
that's awesome
This is really cool and interesting insights, but, a word of caution.
Do not rely on these to dictate your writing. The title is, in my experience, still 50% of the battle, with choosing tags in active categories (without lying) the next 30% and the final 20% being an actually good and shareable article.
The title gets you the early momentum, the "clickbait" gets people in and gets them to vote. An article with early votes then gets algorithm placement and it goes from there.
The tags ensure enough eyeballs see your "clickbait" (by which I mean, an intriguing title, make sure your deliver on what it promises though!), picking a tag with few subscribers sadly does not work well and so this part not feels "gamified".
Finally the content, engaging and interesting (like this post) means you get the shares outside of dev, which normally means Google picks it up early and promotes it on their feeds and that is when an article really explodes.
I spent a lot of time analysing stats and best posting times in the past, then realised the above is the formula that matters (with shares being the biggest driver of growth, but the first 80% is what gets enough eyeballs to earn those shares!).
Hope that helps, signed, a fellow writer who has done this dance!
p.s. I thought of one last thing: keep going! The subscribers count on DEV is way off what I believe are active (or even real) people, but the people who are real and engaged that receive a notification you posted gives you that "early velocity" that feeds the rest of the cascade above. It took me 30+ articles before I started to see that effect, but it certainly does work).
p.p.s. Final bonus tip: comment on other authors articles a lot, they are the most active people here. Leave genuine insights or even arguments against what they wrote, as the active authors are more likely to support you and comment on your work and then that results in even more engagement boost in the algo. Plus, you might even make a few internet friends along the way as there are so few creators compared to consumers it is a great commonality to start chatting from!)
I think you missed the point of the article. This isn't about optimizing for views — it's explicitly about not doing that.
The 50/30/20 formula you describe is exactly the kind of performance theater I'm trying to avoid. My data shows that my most personal articles (fewer views, higher engagement) create better conversations than "optimized" ones.
This tool is for observation, not optimization. Different goals entirely.
Fair enough, I had assumed that the reason to know why someone read an article was leading to trying to find your audience, couple with such detailed stats I had jumped to conclusions.
But there is a difference between optimising that changes who you are and optimising for what gives you MORE of the conversations you are after. I would not call that theatre, I would call that taking the time to play the game we must all play in order to reach more people who want to have the conversations we are interested in.
We live in an attention economy, we must gain attention before we can gain engagement and depth.
Either way I found the fact we could get view times interesting, so thanks for that!
In your finding I think you hit largely why Google built the empire it did. Measuring ctr (clock though rate) is good, but to stay the best they needed the infrastructure that Microsoft had. They needed to see time on site (Google analytics were setup but data should be private). Then a goldmine finding bookmarks in chrome when it first came out were tested as various ranking factors.
I think you're projecting Google's intent onto mine. Google tracks data to optimize ad revenue and search rankings. I track data to understand how my writing lives over time.
Very different goals. I'm not trying to game any algorithm or maximize any metric. Just observing patterns without trying to exploit them.
More just making an observation or a parallel to the first part of your assessment of your analytics. Not all engagement Is good engagement. It is important to recognise this in a lot of fields but I appreciate your deep dive in the first place and reaponse=]