cyle:

this is a really great question, thank you for asking it. you’ve accurately pointed out some glaring holes in our communication strategy – some of that is even my fault, to be honest, as someone who tries to help shape our public communications.

the easiest answer is that there’s a lot of work happening on the internal tumblr side, and not all of it is the same kind of work, and therefore not all of it is being communicated consistently. that’s a huge problem if you’re not inside of it, as you point out; the contradictions seem weird for anyone paying attention.

i’m currently a part of the @labs group and we’re trying to come up with radical ideas for reshaping tumblr, prototype them quickly, show them to people, and iterate on them, and reject them quickly if they don’t make sense. the Tumblr Mini thing recently is an unfortunate example of that – something leaked way too early, before we even got a chance to really understand how the thing we prototyped would be received. that may never go beyond the thing people saw, for good reason: it doesn’t make sense. we’re learning from it, but it’s unfinished.

in that Labs group, we do want volunteer-based feedback, and we’re actually starting that very soon with targeted user research. you may see some surveys soliciting that feedback soon, or invites to be a part of the testing. i’m extremely excited by that, because some other ideas we have feel like they should’ve been a part of tumblr since the beginning.

but…….

there’s a whole different group at tumblr that are making “core product” improvements, and a lot of their work is reflected in the recent staff post about product strategy. their aim is to alleviate a lot of common points of confusion and frustration, some of which will seem counter-intuitive to people who have been on tumblr for many years and learned (what i call) “the hard way”.

a lot of that core improvement work will be rolled out, experimented with, and iterated on, the traditional way: involuntary A/B tests, rather than volunteers. that is standard industry practice, because when you’re trying to understand the behavior of millions of people, just asking for volunteers introduces too much selection bias.

however, it’s important to note that even today’s rollout of the new desktop navigation was released to some XKit devs beforehand, so we could get early feedback. so sometimes we do roll these core things out for feedback first, before “regular users”, before we A/B test it. usually those small initial experiments are hyper-targeted though… in this case, because we do actually care about XKit and third party developers. you don’t see that though!

so full transparency: not everything we do is going to be volunteer-based, not everything we do can be communicated adequately before we test it, not everything during testing will be customizable, because how much we want to customize is dependent on the outcome of these tests. it’s a bit of a chicken-and-egg problem.

and furthermore, this is kind of how it’s always been, for better or worse, since around 2016 when tumblr started believing in A/B tests and experiments, rather than just “ship it and we’ll see”… the common denominator, though, is that we want your feedback. we don’t do a good enough job soliciting that feedback, but we do want it. (i kinda wish we had an easy feedback button for the new layout.) we do make decisions based on feedback, but we augment that decision-making process with hard data from experiments.

i hope that makes sense? there’s too much to cover to provide you with the full context. i could write a book about it. all i can confidently say is that we’re trying as best as we can to balance the ideas of keeping tumblr special and unique, while trying weird/bad/good/uncomfortable new directions to help the platform grow and thrive. status quo and complacency aren’t acceptable; tumblr needs to change to survive. i am just as uncomfortable and upset by that fact as anyone, and i worry every day and night about it, whether the price of survival is worth it.

we’ll see, hopefully together.