Revisiting Agile/UX in 2012

— 10 minute read

John Cutler asked this on twitter:

love the way back machine "In this edition of Ask @uxmatters, our experts discuss the gaps between the agile development model and user experience design." ... 2012 What has changed? Stayed the same?

Here’s the Ask UXMatters article if you want a read.

I started writing a response on twitter… but it got long… so sticking it here rather than annoying all my twitter peeps with a stupidly long thread. This is pretty much a direct brain dump so do bear in mind that this is sitting in the Home Of 💩 First Drafts.

First off… the fact that 2012 is seen as ye olden days makes me feel ancient :-) It's seven years ago. I own shirts older than that.

I'm gonna attempt to answer this from my POV. I've got some pretty darn obvious biases (UX & Agile are peanut butter and jelly as far as I'm concerned — they go even better together :-)

For framing folk should understand how Janet builds these articles. The question gets emailed out, folk reply, then Janet Six edits everybody’s ramblings into vague coherency. So you don't generally see discussion or comments on other people's answers — since folk aren't in the same room to discuss :-)

(The above turned out to be an oversimplification — because Adrian is an idiot who assumes rather than asks — according to Pabini Gabriel-Petit "Janet assembles everyone’s responses, writes the intro, adds headings, & often writes her own responses, then I edit everything & add my own responses, if any. So there are 2 of us who respond to what others have written." :-)

Let me try and set the context of 2012 a little in Agile/UX relations from my POV. This is coming from memory and I'm currently too lazy to do serious fact checking. So some of these dates may be a tad off. Let me know.

It's 2012.

The Agile Manifest is 11 years old. XP, Scrum, DSDM, etc. are even older. Agile Alliance founded in 2001. Agile had crossed the chasm from the early adopters. The tediously annoying (K|k)anban vs (A|a)gile vs XP vs Scrum turf wars are beginning to settle down. Most people have heard of agile. We're in early majority territory.

JJG's "The Elements of UX" diagram is 12 years old. 19 from Norman inventing the "User Experience Architect" job title. IxDA & IAI both founded around ten years back. The tediously annoying Big IA vs Small IA vs IxD vs Usability turf wars are settling down. This is the year UPA rename themselves to the UXPA. Most people have heard of UX.

(As an aside it wasn’t until I wrote those dates down that I realised that "UX" as a popular label and "Agile" as a popular label have kinda-sorta the same timeline and hype curve. Interesting that…)

There have been folk having focused conversations about UX & Agile almost since the terms came into common use (shout out to the old agile-usability mailing list, started around 2004 and soon to vanish in the Yahoo! groups purge.) There has been a dedicated UX track at the Agile 20XX events since 2008.

Steve Blank's first customer development book had been around 7 years. Which begat Lean Startup soon after. Smart folk like Janice Fraser started talking about smooshing lean startup & UX together almost immediately. I'd first heard about "Lean UX" via her work at the 2010 London Agile/UX Retreat. Jeff & Josh's Lean UX book is still a year in the future so most folk haven’t come across the concept yet yet.

So… let's revisit some bits of that UXmatters piece with that context…

"UX design, we want to know everything that’s relevant up front"

Since smart folk like Desiree Sy had been talking about doing doing discovery & dev work in parallel for five years at this point this certainly wasn’t the only view at the time.

Desiree Sy’s original piece on integrating discovery and development work together got labelled dual track, and then repeatedly misinterpreted as saying best practice was to have separate dev & discovery teams. This is wildly annoying. Jeff Patton has a nice piece on how this happened.

Since then the popularisation of Lean UX, Lean Startup, Assumptions / Hypothesis / Experiments, Outcomes over Outputs across dev/ux/product have all pushed folk further from the whole UX own/articulate the vision thing. We’ve all been finding better ways of articulating this stuff and building artefacts that everybody can align around.

"I volunteered to write the user stories myself." / "even if it means that I need to take a greater role in requirements definition than a UX professional typically would"

Something I notice reading this now is the almost complete lack of any mention of product / product management. Some of what we see here is folk stepping in to fill gaps because there was not mature product focus within the organisation. Product as a community of practice has been growing, maturing, and defining itself much more strongly over the last ten years. These days we’d be more be seeing product managers and product management popping up more often in these conversations.

"When UX instead resides in a services group, it’s often a struggle for UX designers to get developers to think of them as an indispensable part of a product team."

Yup. That separate UX group working by itself on its own thang still causes problems. See it still, but less often than I used to. The flip side of that is the dev group that won’t let UX in. Likewise — see it still, but less often than I used to.

This has pretty much nothing to do with agile/ux. The same sort of stuff was happening in the 80s and 90s. It’s a people/org problem.

“In my mind, the best way to combine agile and UX is to use Sprint 0—which, in an agile development process, is when you define the backlog—to do UX design, which then feeds into the backlog,”

Hated Sprint 0 then. Hate it now. Well… hate is maybe too strong a word. But this wasn’t a hugely common agile practice, and was originally mostly about how we get a new team up and running, setup the software, source control, etc. Not about defining the entire darn project.

But the folk new to agile needed somewhere to slot the "requirements" bit of their stage gated development process. And "Sprint 0" sounded like a good place for that. Which sucked from both directions.

People either came in with something so well defined folk were chopping up a requirements doc into small chunks — rather than iteratively solving customer problems. The other extreme being folk kicking off with no real idea of the end-users, the problem, the potential space of solutions — where dev work kicked off too early leading to lots of waste.

Annoying as all heck.

Sprint Zero. The "Design Sprint" of its day.

"I find that UX does a good job of identifying and designing for the happy path, but in my experience, the main gaps have come around the alternate paths and exception cases that did not emerge,”

Yeah that happens, and still does happen on occasion.

The complete opposite also happened, and still happens on occasion — where UX do a good job of talking about stress and edge cases and dev doesn’t build 'em.

Both of those instances suck — have everything to do with the people and teams involved, and nothing to do with agile & UX.

"Writing User Stories"

Huge arguments about terms, formats and approaches to this then and now. Other approaches to organising stories — like user story mapping — have already been around for a good few years at this point. Lots of people talking about ways of writing user stories that aren’t "as a user…".

That strongly hierarchical breakdown of theme/epic/story is something that rightly gets a lot more push back in some quarters these days. I think we have better ways of articulating product direction now. Actually we had them then to, but they’re more talked about now :-)

People aren’t always trying to mash everything into epics/themes/user stories and backlogs these days. We got assumption maps. We got user story maps. We got OKRs. We talk about assumptions and hypotheses and experiments. We talk about vision and strategy deployment. We have kanban boards on the wall that show where discovery is happening, what options we’re trying to explore, and how we’re trying to explore it. All good stuff.

What I do see less of — thank goodness — is that two inch thick requirements document folder.


Other things that occur about the difference between then and now.

  • A lot of that piece reads to me like people in a project rather than a product mindset. There’s a start, there are some iterations, and an end. So there’s that focus on forming a vision and on requirements. That stuff gets a different focus when we’re in environments that are continually delivering and evolving products. That’s not to say folk weren’t thinking about it back in 2012. But there are a lot more people in that mindset now.
  • The Lean Startup / Lean UX stuff has gone from fringe innovator / early-adopter groups to vaguely mainstream. Let’s say early majority time. It’s much easier to talk about these things now. The conversation around assumptions / hypothesis / experiments gives us a conceptual framework that crosses business/product/ux/dev divides and is something I’ve found super useful.
  • It also means there are a lot more people doing a terribly bloody job of it. Because that’s what happens when stuff gets popular.
  • The user research voice is fairly absent in that piece. For me it's been getting louder in recent years. Especially (from my POV anyway) after the boost it got from the custdev/lean startup hype boom. That reframing of the importance of talking to and understanding the end-user and customers has helped me surface the importance research across the org.

I should shut up now and go do some useful work while there is some of the morning left.

One last thing before I go.

I’m slightly weirded out by how adoption and change feels simultaneously slow and fast.

We live in a world where some people have never worked in an environment without a continual discovery/delivery mindset — and anything else feels alien and bizarre.

At the same time other companies and teams live in a world where delivering working code once a month is a pipe dream, have product managers that just ship feature requests from customers to dev teams, and have three designers trying to serve hundreds of people.

The dissonance between those extremes is intense. When the latter try and adopt the former's practices odd things happen. They're so far apart it's hard for either side to even hear what the other side is saying.

You see some of that spread in that 2012 piece. You see that spread now. We keep talking about "UX" and "Agile" and "Product Management" (and whatever) as if they're monolithic things with fixed definitions. They're not. They're Reifications They're a messy fuzzy spread of people, practices and context. Containing mutually inconstant definitions and labels. Walt Whitman's "I am large, I contain multitudes." line springs to mind. How we see our communities and how our communities actually are can be very different.

So… y'know… try and be excellent to each other.

ttfn.