Speed vs accuracy: the fundamental challenge for process design
Prioritizing correctly is crucial
A joke I recently read highlighted the fundamental trade-off between speed and accuracy. The joke is presented below. The trade-off though is nothing new. It’s something we all know and have experienced it, directly or indirectly, across a variety of areas.
It has important implications for process design. The prize for getting it right is significant. The cost of getting it wrong is a long slippery slope.
Trade-off — with a nuance
The joke goes (somewhat) as follows:
- Person A: My maths speed is really fast 😀
- Person B: Really? So, what is 371–129? 😐
- Person A (replies in 1 second): 94. 😅
- Person B: That’s wrong! You said you were really quick at maths?! 😦
- Person A: I said I was always fast — not always correct 😜
While this trade-off is fundamental, in reality it’s more nuanced than the humorous example above.
It’s composite and relative.
Up to a point, speed and quality are composite. They are co-related until the point a process offers its baseline functionality. After a certain base performance is achieved, the trade-off becomes relative. Past this point, the trade-off between speed or accuracy must be acknowledged and addressed.
Interventions
A classic pitfall is having an arbitrary, blanket or unquestioning focus on one aspect over another.
In one organization, many processes were arbitrarily optimized purely for speed. Speed of transaction was taken as the holy grail, often to the detriment of quality. Any perceived impediment to speed was instantly dismissed as draconian, without understanding its rationale. The need for speed was paramount. On top of that, the interpretation of speed was worryingly narrow — purely based on step count. The fewer the steps in the process, the better. The resulting processes were fast but of course hugely error-prone — much like the joke presented earlier.
Another example took the other extreme. Here excessive checks were put in place. The purpose was noble and the team had rightly focussed on accuracy. However, despite over 99% accuracy, which for a manual process was fantastic, the process was slow. The team wanted the best result for the customers but that came at the expense of (avoidable) overprocessing and staff overtime.
Both cases were fundamentally similar. Both had an overly narrow focus.
To be sure, in some cases it is required to take sides — and decisively. In some cases, quality is an absolute even if it comes at the cost of speed. This could be driven by marketing concerns (differentiation) or by safety concerns. The latter is more often the case. Equally, in some cases, speed is absolute and like above, driven by marketing or operational concerns.
When there is a (sound) strategy, a core operational concern or an overarching mandate, that is a valid reason. In other cases, it’s important to critically analyse the implications of being along various points of the speed-accuracy continuum. It’s particularly important in processes where there are is a considerable avenue to juggle speed and accuracy.
Critical analysis is important.
At a conceptual level, it's about selecting guiding principles which address this trade-off with master strategy. At the task level, the challenge is to be able to optimize various parts of a process for speed or quality to achieve the best overall outcome.
Striking the right balance
Speed is a tempting aspiration. It sounds good and is easy to sell. “The process must be fast” — is hard to argue against. Speed is important. So is quality. In fact, up to a certain point, speed and quality are both important (and expected). After a certain point, the trade-off kicks in, which requires designers to prioritize one over the other.
But what do we go for? Borrowing from LEAN literature, it’s all about value. This is the fundamental starting point — understanding what value means — in a particular context.
A process is a means to add value. Every process will have a different value proposition. Depending on that, we can decide (or derive) an optimal point in the speed-quality continuum.
To strike the right balance, it’s crucial to understand value. It’s important to understand what is most important in a given scenario or situation. This assessment must be done holistically. It must balance outputs against outcomes and consider the value stream in its entirety.
There is the concept of time to value. Ideally, this must be a short as possible. This heuristic, however, can be misleading if not understood in context. The defining term here is value. If value is compromised, time to value loses meaning. Quality (which includes accuracy) is implicit in value.
The first step is achieving a baseline quality level which produces a robust output that fits the customer’s purpose.
After that, time to value ratio can be progressively shortened by systematically eliminating non-value-added steps.
Time to value is shortened by reducing non-value add steps, and not by an indiscriminate reduction across all steps.
There is no shortcut to getting this right. The proven path involves going to the Gemba, or in other words, experiencing a process first hand (ideally) or analysing it deeply to understand the real picture. In design thinking, this step is called gaining empathy. The broader and deeper our understanding, the better we can address this trade-off.
Speed vs accuracy is an inevitable and unavoidable trade-off that process designers have to address. However, by understanding drivers of value, and the ground realities of their context, it’s possible to prioritize correctly, and then, via a series of tactical interventions, ensure that a robust balance is struck between efficiency and effectiveness.
