I feel a much longer post coming on about the topic of “optimizing organizational models”, but I wanted to share a few quick thoughts on the link between organizational models and innovation. Given the recent events at Yahoo, there has been much discussion about why the company continues to struggle and what can be done about it. The infamous Yahoo matrix organization is one point that is raised often.
From a recent post by Salim Ismail on GigaOM about Yahoo’s matrix:
But it is terrible for accountability or speed. Whenever you launch or change a product, you have to get clearance from legal, PR, branding, privacy etc, which inevitably takes time. The matrix structure also prevents any real risk taking. The legal department, for example, wants the same ToS across all the products. Brickhouse was created to circumvent this issue. Set up outside and away from the mother ship, we hoped to be the tugboat that pulled the big tanker around. It worked for a while, but the Microsoft bid pretty much derailed that effort. The company had to focus all its energies to fend off the bid, which was necessary but incredibly disruptive to morale and productivity. In addition, the Matrix had woken up and was attacking our unit (the best analogy I’ve found for this is that whenever you do corporate incubation, the immune system of the company will come and attack you — but that’s a whole other post).
The matrix structure works great in older, slower industries, but on the Internet, the two attributes you must have are speed and risk. Very simply, Yahoo’s organizational structure is antithetical to the industry they’re in. Over time, that structure has calcified and today, Yahoo is a 14 year old dinosaur in the industry it helped form.
I agree that the matrix only makes sense for certain support functions that you really should centralize, usually due to policy and oversight (e.g., HR, Legal). But, pushing other functions like Design and Research into a matrix role tends to slow down execution and hinders innovation. Decision-making becomes fuzzy when different team members actually report into different organizations. Accountability and “loyalty” are also unclear when someone works day-to-day within one organization, but has a manager outside of that organization who may have goals that are not fully aligned with the goals of that product team. I know, because I have experienced this personally as organizations decentralized into a matrix role and then re-centralized again, several times over the period of just a few years in a number of corporations where I was employed. The cycles of reorganizations and changes in ownership, decision-making, and accountability all contributed to execution issues.
However, I’m not sure that I would characterize the Brickhouse team at Yahoo as executing successfully until the Matrix “attacked it”. Everyone within Product and Engineering organizations should be responsible and accountable for the innovation and evolution of their respective products. Creating a separate “innovation group” that exists somewhere else in an ivory tower only makes sense if that group is chartered with the creation of an entirely new product. There are historical examples of that having some mixed success. For example, Apple’s original Macintosh team delivered, but the rest of the company felt alienated. If the innovation is supposedly being applied to an existing product line, you can be certain that the core team will feel alienated and exhibit “organ rejection” when you try to inject that innovation back into the core product. Also, this separate innovation group typically does not have the deep knowledge and experience with a product that the dedicated team does, after years and years of design and development. Yes, yes. I know that a team can stagnate and fail to innovate their own product because they are too close to it. But, that can be fixed. There are many ways to remedy this through intelligent hiring, better organizational models, and sustainable changes to product process that help drive repeatable innovation. Thinking that you can create product innovation in a vacuum, without collaboration with the existing team, and have it succeed in the market is a myth.
It simply does not work.
More to come…