Today I'm moving Yupdates into public preview (!).
The 9+ month private preview period was exciting and rewarding (and, at times, frustrating). I received great feedback that changed the course of the service, and I'm grateful for everyone and their generosity in dealing with rough edges.
Yupdates is a new take on feeds. It's usable by anyone but geared toward people in the software industry.
It's focused on something few products do well: triage.
There's always going to be more out there: more posts, more tweets, more toots, more tutorials, more products, more releases, more drama, more nonsense, and, thankfully, more cute pet videos.
Social media has warped much of this into controversy/outrage gaming, but I'm still optimistic. There's new, high-quality content posted every day, and for work, there's new critical information out there every day. We just need to be able to spot it (without wasting hours and hours).
Time and again, the tools and algorithms that try to show us the most interesting or important things have come up lacking. We're complicated, and our work is complicated.
They also jumble everything together; when we're online, steeped in a particular context, it's better to stay in that context (whether that's our job, a specific project, a subject we're studying, a hobby, etc.).
Embracing triage means you point in a general direction — "these 100 sources often contain important or interesting content about XYZ" — and look through the options when you're in the "XYZ" context.
I'd be happy if all you leave here with is that thought, and you take it to your product designs and customer conversations. Our email clients, social media apps, wikis, feed readers, and project management tools can all be so much better if one of the "jobs to be done" we all care about is triage.
What are the key ingredients? To start: many results at once, the ability to skim, keyboard shortcuts, and very few steps to get the app in line with our next thought. It's all about flow.
Search is essential, but with new content, it's more often "I'll know it when I see it." We can only provide search keywords for things we've heard of before.
Automation can be helpful (like an "if X, always do Y" rule), and Yupdates has a system for that. But an under-served way of working is something we might call "half automated." There are so many things we need to stop and think about, even if we only need a few seconds.
Is that post relevant to work? Is that an email I need to deal with? What tags should this have? Do I want to read that later (or now)? One of our dependencies has a new release - is there anything urgent? Is this GitHub issue assigned to the wrong person? Does this post about our product need a follow-up?
After making a decision, we may want to take an action. This could mean tagging an item's priority, adding it as a to-do item, reading it on the spot, emailing it to ourselves, bookmarking it somewhere, or triggering a POST to a custom workflow. Or (in most cases), just ignore it.
To triage well, you also need performance and good information density.
And you can't fake performance. It needs engineering skill and operational excellence, all on top of a good design and sound cost model. There's no pixie dust we can add to existing products to solve that problem quickly. Of course, customers like a fast app, and most product managers realize that. But I keep running into products where, even if it isn't dog-slow, they lack high-level browsing views and assume I find each item equally important.
With nuanced topics — like our careers and deep interests, which we pursue for decades — we'll judge what is relevant or interesting (thank you very much).
Another big focus in Yupdates is rich, varied input sources.
Everything's not going to conveniently conform to one format—ever. RSS started over 20 years ago, and it remains popular in many circles (and Yupdates supports it, with love). But stopping there is limiting; there's valuable data in so many places. With custom integrations, we not only get access to more, it may even be better. We can offer useful, new semantics.
For example, Yupdates tightly integrates with GitHub, calling their API regularly, creating a custom stream of issue events, pull request events, releases, or notifications for you. Because the object model for each event is not an opaque blob of text, we can offer filters based on the type of event (see the GitHub demo and this FAQ for examples). You can combine that with the universal text filters to zero in on the areas of a project you're responsible for and more casually follow the rest (or not at all).
That is much better than sifting through hundreds of notifications and emails, clicking each one to find out if it's even relevant.