“The most general maxim for those who study functionally organized systems is that we come to understand how things work by studying how, when and where they break down.”
— William C. Wimsatt,
Re-Engineering Philosophy for Limited Beings: Piecewise Approximations to Reality,
p. 22, Harvard University Press, 2007
What does this quote from an academic who studies evolutionary biology have to do with business?
As it happens, plenty.
In business, how we think things work can be thought of as a system or systems. But often we use terms such as models, processes and so forth, instead of system.
So, let’s consider how models work…
We think we understand how “things” function, either just in our head, a company process, or expressed as a series of calculations (think Excel). We use models like this all the time. In my world, financial models are very deliberate and explicit, but we also have models of how we think people individually or collectively will act, how customers think and buy, how an industry works, and so forth.
So when “things break down,” what that really means is that something didn’t turn out the way we thought it would
— our model (or process, or understanding) has some flaws and, therefore, can be improved. Understanding why things happen differently than expected can be important.
Using breakdowns productively
There are many ways to use breakdowns to improve a business.
One is to tap into
small breakdowns as a means of driving steady increases in understanding.
For example, when I help companies with acute liquidity problems, I develop a model to forecast cash flow on a weekly basis. Tracking and comparing what actually happens each week means I am digging into the “breakdowns” all the time. That allows me to quickly improve the model and my understand of how that client’s finances really work. If time allows, I go through several weeks of iterations of these models before locking them in with creditors in forbearance agreements and so forth.
In less liquidity-strained companies, I like to implement a monthly forecast/actual cycle.
Again, a financial model is built and based on operational metrics, etc. Then, each month, we look to see where the model “broke.” Doing so allows us to fine-tune our understanding and also quickly spot when key assumptions change, such as customer sales levels, material prices and so forth, not to mention the common assumption of good execution on our part.
An analysis of breakdowns can also have a big impact on profitability or other desired outcomes.
For example, one client of mine makes custom specialty chemicals in batches for its customers. Standard quality assurance measures have the technical department test each batch for matching the spec in terms of whatever the customer wants, be it viscosity, color and so forth, before final packaging for shipment. Since adjustments are sometimes needed to meet the specification, the company tracks how often the batch passes on the first test and has begun a project to understand why this isn’t always the case. Here, my client will be using some of the standard quality techniques, such as root cause analysis and others.
Another example is when a deep dive is inadvertent and complements an incremental approach.
Years ago, when I consolidated seven accounting centers (some of which were empty and not running at the time) into one, we knew we were diving into an accounting mess. One tool to fix this is the monthly process of reconciling general ledger accounts to source data. When things are not going well — “as expected” — there are lots of variances that have to be explained.
In my situation,
there were too many breakdowns to explain at once.
So, my people kept closing the books monthly and working down the list of reconciling items which could be explained. In preparation for a balance sheet review (where all of the account reconciliations are reviewed in mind-numbing detail at once) by the Corporate Controller, I held a dry run with lots of accounting staff present, from all of the various departments.
This dry run ended up being a deep dive: all of a sudden, patterns emerged,
as “matching” reconciling items could be seen in different accounts. This led to discovering lots of glitches in subsystem feeds.
Breakdowns can also be useful in debriefing new product failures.
Uncovering what caused a variance between expectation and reality can improve the understanding of customers, competitors, the market, and more.
Overall ideas for putting this into practice
- Before spending the resources, make sure there is a perceived benefit from improving your understanding of how things work. Don’t study every failure, just the ones that are critical or otherwise matter.
- Keep in mind that in many cases, repetitive analyses (as described in the financial projection examples above) will result in more progress than a single, big, deep dive.
- In concentrated areas of failure, a deep dive is often the best approach.
- Don’t forget the role of people in all of this. They fail too, and it is important to understand why.
And finally,
what if rather than being surprised, you simply have no idea how something works?
In that case, use one of Wimsatt’s tricks: Build a knowingly flawed model and then start figuring out why it doesn’t work. Now you have breakdowns, even if they are contrived.
In business, we are all invested in seeing things succeed. And rightly so. That said, when things “break down,” as disappointing and costly as that may be, there is often opportunity and wisdom beneath the surface.
Investigating these breakdowns can lead to significant and lasting improvements.
