Why Designing Systems is Difficult

A little while back I wrote that my new mantra is “The system is the product”. The context of this was that every product today, even something as simple as a kitchen table, sits within a system of meaning, distribution, retail, service and brand. The system can be as important to the success of the product as the product itself, and can have a massive impact on the customer experience of the product.

I didn’t really expand on it much in the original post, but I actually got a lot of good feedback on it. Peter Merholz was nice enough to pick it up as part of his talk entitled “Stop Designing Products” at the SHiFT conference in Lisbon (download the PDF of his excellent talk here).

One of my first experiences with tackling a big systems problem came a number of years ago when we had a client at frog design who was seeking to develop a car-sharing business with a custom-developed electric car. They had made significant headway on the engineering of the car itself, but other elements with far reaching impacts on the business had been largely ignored so far despite the fact that they were critical path on the schedule: Defining who the customers would likely be, how they might use the service, how things might go wrong with them renting an unfamiliar electric car, how the end-to-end service experience would be (signing up, making reservations, returning the car), how to load-balance the fleet across the city, how to manage maintenance, and so on. These were all much more abstract questions with much less black/white answers than selecting the right battery vendor for the car’s powertrain or how to ensure it passed crash safety standards. The product was getting all the attention at the expense of the enabling system.

Systems are really hard to do well, much harder than individual products. There are several reasons for this:

  1. They are abstract and hard to see. Products tend to be tangible (whether physical or digital, you can “get your arms around them” conceptually or literally). This means systems are hard to understand and therefore difficult to work on. The flipside of this is that customers tend to see the system through the interface points, as Dan writes at Adaptive Path: “What users physically experience represents the system to them, and how it works. The interface is the system.”
  2. Systems are the undervalued connective tissue. As my colleague Luke Williams argues, we are at a point in advanced capitalism where it’s the connections between products that are more important than the products themselves. But the value of this connective tissue is consistently undervalued in our culture. Just as we value specialists over generalists (surgeons are paid a lot more than family practitioners, for example), we tend to like things that are easy to define, not diffuse things that aren’t, even if those diffuse things are vital to the proper functioning of the easily defined things.
  3. Systems are often boring. They are not shiny and interactive, and their features tend to enable the things they are tying together, rather than being obvious features of the system itself (in fact if the system is doing its job properly it should largely disappear).
  4. Systems cross over organizational boundaries. This is because the customer experience they are enabling crosses over those same boundaries. This makes them an administrative nightmare, and bubbles up all the tensions, insecurities and divergent directions within an organization. Getting systems to work well and come alive often takes a certain dictatorial spirit. There’s a good reason that Apple makes good systems.

We have become very good at managing product development. We now need to get much better at managing systems development.