This morning, Dan Saffer shared on Twitter that Smart Design SF is shutting down. Earlier this month, Adaptive Path surprised pretty much everyone who cares by announcing they were being acquired by Capital One. So what’s going on?

I don’t know the rationale for Smart’s decision, but in talking to friends, and trying to make sense of what’s happening, it appears there are two opposing forces that San Francisco design firms have had trouble reacting to.

The first is the growth of well-paid in-house design teams at money-hoarding tech companies and well-funded startups. I first saw this about 10 years ago, when Adaptive Path lost a candidate to Google because they offered what we thought was an insane salary. A few years later, Facebook and Twitter followed suit, and now every tech company is offering designers 50-100% more in salary than what design agencies can swing. In the past, agencies could say, “Yes, but we respect and value design in the way that in-house companies don’t, and you’ll get to work on a range of things, instead of just one thing over and over.” That doesn’t hold true anymore, and most of the interesting design work is emerging in-house, and designers want to be where the action is. And get paid better to boot.

Here comes the counterintuitive bit. If design is in such demand (and jobs pages at every company suggests it is), why aren’t agencies just charging more, and using those higher billings to pay their designers more, and thus be competitive? That’s how markets work, right?

Well, not quite. What’s actually happening, according to friends at agencies, is that client’s willingness to buy design from agencies is decreasing, and project budgets have been shrinking. And the prevailing theory is that this is happening because companies are building in-house teams, and that’s where their ‘design budgets’ are going. Whereas in the past, a company might spend 20% of a design budget internally and 80% externally, that’s now swapped.

Another way to put it, as companies have gotten smarter about design, and recognized it’s a competency they need internally (something which Adaptive Path promoted in our book), they’ve become less comfortable outsourcing it.

So, San Francisco design agencies are often billing less than before, yet the talent market is able to earn more than before. This is a quandary.

It’s telling that Smart Design seems to be otherwise financially healthy — last month they opened an office in London. Their main offices is in New York. It just seems like they’ve thrown in the towel in SF, and from what I can see — I don’t blame them.

When presenting my philosophy of design organizations, I show this slide:   On the left, comprised of 5 robot lions (stay with me…), you have super robot Voltron, Defender of the Universe.  On the right, 5 unicorns. The unicorns refer to the “design unicorn” a designer skilled at interaction design, visual design, and front-end development. They’re “unicorns” because some believe they don’t exist, though others are betting that they are very much real. For the sake of argument, let’s say that design unicorns are real.

My concern is that all this talk about unicorns, on thus on design at the level of the individual, is misguided, and potentially even harmful to design as a field.

Companies want a unicorn because she can be the lone designer on a product team, working with product managers and engineers to deliver features. On the face of it, this seems great – you get a lot of bang for your buck.

However, in my experience, design is best considered at the level of a design team, and the point of the slide above is that a coordinated team of 5 designers (of varying skill sets) can out perform 5 identically-skilled unicorns working in isolation.

When I joined Groupon, product designers primarily acted on their own in product teams. And what I saw is that the scope of their efforts was quite constricted. When you’re asking one person to deliver interaction design, visual design, and front-end code (or at least something prototyped), it leads to designers focusing on narrowly defined features — a single web page, or augmenting an existing flow.

Among the first things I did as VP of Design was relax our requirements for product designers. They no longer needed to be technically savvy, and, they could have an emphasis in interaction design and IA or in visual design.

And much of my own initial effort was on hiring strong design leads — directors and senior managers. This turned out to be key. About 9 months into my tenure, I organized the 40-plus person Design Union (what we called the whole design team) into a series of teams, each with a lead. This proved transformative not just for design, but the for the company. With these teams in place we could deliver on larger scale efforts like Groupon’s first-ever website redesign (a prior effort, before I had joined, had failed), or DealBuilder, a self-service tool for merchants to launch a deal.

Another short-coming of focusing on design unicorns is that it reduces design to an execution function (I’ve written about definition and execution in my post and presentation on The Double Diamond.) Approaching design from the orientation of teams, with strong team leadership, not only allows designers to tackle broader problems, it also enables them to move “upstream” into the decision-making behind product definition.

Team Structure

Through experience, I’ve arrived at a sense of what seems to work for the structure of design teams (again, these are teams within a larger design organization, not the whole organization itself!).

Teams should have 4-7 members. Any larger than that, and team coordination gets unwieldy.

Teams require strong singular leadership. This Team Lead role is perhaps the most challenging in the whole design organization (even moreso than a VP of Design). The Lead must:

• Manage down – getting the most and best out of their team
• Manage across – collaborate with cross-functional partners
• Manage up – presents to executives and other stakeholders

Pivoting between detailed delivery within your team, coordination with peers in product, engineering, and marketing, and communication to executives can induce a kind of whiplash that not everyone can handle (in fact, I would say it’s rarer than the ‘design unicorn’). But if you find folks who can handle this successfully, your design organization can be surprisingly effective.

These teams should be organized around business problems (supporting a line of business) or aspects of the customer experience, not by function (interaction design team, visual design team, etc.)

And within these teams should be located a spread of skills — user research, strategy, planning, interaction design, information architecture, visual design, and prototyping. And this suggests another reason to Think Beyond Unicorns, as no one could be expected to excel in all of these areas. Instead, hire folks who excel at a couple of these but can productively collaborate across all of them. (I’m not in anyway advocating the hiring of single-minded specialists.)

There’s an alchemy that happens with great design teams, where the whole is greater than the sum of the parts. My understaffed teams were able to deliver far more than could have been reasonably expected because they could leverage their team-ness for greater effectiveness.

When I next write about teams…

Something that I’ve realized is that the ideal level of scope for a design team is different than that of a product and engineering team, and designers are currently constrained by working within an organization optimized for the latter. When I have time (and after I do some doodling), I want to talk about bridging that gap.