03 October 2011

more agency start-up cross-pollination

So, in part I of my (slow developing) series on what Agencies can learn from Start-ups, I introduced a whole bunch of issues, lessons and models.

As promised, I'm going to start expanding on some of them.
First - in an agency it is not okay to fail. Failure is not an option, ever. In a start-up, on the other hand, failing is how you learn and how you get better and how you end up winning. Learning from your failings is, in fact, built into the way you do work. By building a culture and a model around an unwillingness to fail agencies not only massively decrease the odds of doing something interesting and innovative - they also massively increase the cost and effort of doing anything at all.
So... how can Agencies learn to be more willing to fail?

To me there are two fundamental issues at play here.
Agencies first of all live and die by account growth. Having to acquire new accounts to replace lost accounts is incredibly expensive and painful - and new accounts always yield far lower profits than existing ones. This tends to breed risk averse behavior.
And secondly, most Agencies (regardless of what they say) are still executing on the basic "single chunk of discrete work with a due date" model that started in traditional print advertising. This creates a viewpoint that a piece of work (a "project") is discrete and time bounded and one day will be "done."

So Agencies look at the work they do and say, "it needs to be perfect and done on-time to satisfy the client."

Folks in start-ups tend to shake their heads at this entire idea.

So what do we want to see Agencies behave like in this case?

Well... first of all, instead of "pixel perfect and on schedule" they need to get to the point where work is not "done" ever and where "perfect" is far less important than quick.

Secondly, a start-up would understand that this kind of work is experimental, exploratory and high risk and would have built an approach that assumed a high probability of (initial) failure and would have assumed that there would have been multiple attempts with multiple rounds of (public) change in order to gather data as quickly as possible and then adjust direction based upon this data. Agencies, meanwhile, tend to arrive at a conclusion and build it as if it were going to succeed.
To make this work, Agencies would need to do two things.

First - they would need to fundamentally alter the way they pitch clients and work with clients. Instead of the "we're fucking amazing experts" pitch - they need to be willing to say "we're really good at working with you and your customers to figure out the best solution." They need to structure work so that there are a series of releases (private alpha, private beta, beta, etc) and with each one build in the ability to gather data from the users to inform decision making on the next release. They need to collaborate with the client so that they can make quick data driven decisions.

Second - they need to change the culture of the Agency. The sort of ego-driven and highly control-centered culture of an Agency is hostile to things like data-driven decision-making and actually listening to customers and users. Rigid waterfall project plans designed to cover the Agency's ass and enable endless (profit generating) change orders do not work in a scenario where incremental releases generate data which then generate insights -- which then entirely reform your project and product plan.

So... yeah... sounds pretty unlikely I know.

To me, there seem to be two ways Agencies could make this actually work.

Either by creating a new Agency from scratch designed to do this kind of work in this kind of manner - or by blowing up an existing Agency and bringing in new management with a broader background.

In both cases, the litmus test is less likely to be the clients won or the work produced initially and more likely to be the talent attracted and the strategic partnerships closed. If you do this right - all of a sudden an Agency is going to have (for example) engineering talent that no other Agency can rival. And the strategic partnerships (with folks like Twitter or Bitly) are not going to be the on-paper PR deals that we see now and instead a true, reciprocal partnership.

Normally I'd dismiss all this as a pipe-dream of limited value.
But here is the thing... in looking at this you can see a model for a new kind of agency - an entrepreneurial agency - that not only does things smarter, but actually makes more money. And most of all... which makes their clients far happier with the work than what we see at present.

I can dream can't I?