The observation
An approval process that worked adequately when it handled thirty requests a month is now causing visible delay at ninety. Requests pile up, people complain, and the process — previously unremarkable — becomes a recognisable source of frustration. The most common conclusion leadership teams reach is that the process itself needs streamlining: fewer steps, fewer sign-offs, a faster route from request to decision.
That conclusion is plausible, and sometimes correct. It is also frequently only half the story, because volume is rarely the only thing that changes as a business scales. Growth changes who is submitting requests, how familiar they are with what “good” looks like, and how much context they bring with them — and all of that can shift the actual work involved in each approval even while the process itself stays identical on paper.
What leaders usually conclude
The instinct to streamline is understandable and grounded in something real: a process built for thirty transactions a month genuinely may not be well designed for ninety. Processes carry fixed overhead — a sign-off step, a review meeting, a queue — and fixed overhead that was tolerable at low volume becomes a larger proportion of total time as volume rises. This is a well-established feature of process design, and in many cases the fix genuinely is structural: remove an unnecessary step, parallelise what was sequential, or delegate authority that used to sit with one person.
What this explanation tends to leave out is a second variable that often moves in the same direction as volume: the quality and completeness of what is arriving at the process in the first place. As a business scales, the people generating requests are frequently less experienced with the process than the smaller group who used it when it was new — newer joiners, people further from the original context in which the process was designed, people operating under more time pressure themselves. If a growing share of what arrives is incomplete or requires clarification before it can even be assessed, the process looks slower not because the process has degraded but because its inputs have.
What the evidence suggests
Consider an organisation where exactly this ambiguity plays out. The approver overseeing the process reports two things simultaneously: utilisation is high — the process is genuinely busier — and a growing share of incoming requests arrive incomplete, requiring back-and-forth clarification that was rarely necessary when volume was lower. Two variables have changed together: how many requests arrive, and how ready each one is to be assessed on arrival. Their relative contribution to the visible delay is not obvious from the outside, and — this is the harder point — it may not even be knowable without deliberately separating the two, because the compounding effect makes each one look worse than it would in isolation. A modest rise in incomplete requests, layered onto a genuine rise in volume, can produce delay disproportionate to either cause alone.
This distinction matters because the two explanations point toward different territory. If the primary driver is genuinely structural — a process with fixed overhead that no longer amortises well at higher volume — the relevant question is about the process’s design. If the primary driver is input quality, the more relevant question is upstream: why are requests arriving less ready than they used to, and does that reflect something about how new joiners are oriented, how clearly expectations are communicated, or how much context has simply been lost as the organisation has grown past the point where everyone submitting a request already understood, informally, what a complete one looked like.
Larry Greiner’s research into organisational growth offers a relevant frame here: processes and practices that work well in one phase of growth are calibrated, often invisibly, to the conditions of that phase — including the shared understanding that exists among the people using them. As an organisation scales, the population interacting with a given process changes faster than the process itself does, and some of what made the process work was never written into the process at all; it lived in the users’ prior experience. When that experience is diluted by growth, the process appears to fail, even though nothing about its formal design has changed.
There is a genuine trade-off in how leadership teams typically respond to symptoms like this. Streamlining a process is visible, concrete and satisfying — it produces something that looks like action. Investigating whether the real driver is upstream, in how prepared people are before they reach the process, is slower, less visible, and can feel like it is not addressing the complaint directly. Both responses can be legitimate. The risk is committing to the first without having genuinely ruled out the second, because a streamlined process still fed poor-quality inputs will likely produce the same frustration in a different form.
There is also a contextual dimension worth noting. Organisations that grow primarily by adding volume through existing, well-understood channels — more of the same customer type, more of the same request type — tend to see the input-quality effect less acutely, because the population feeding the process has not changed in kind, only in size. Organisations that grow by adding new segments, new geographies or a wave of new hires unfamiliar with prior informal norms tend to see it more acutely, because a larger share of what arrives at any given process is now coming from people who never absorbed the unwritten context that made the process efficient in the first place. Two businesses experiencing the identical symptom — a slowing approval process — may therefore be facing genuinely different underlying situations, and the same fix will not necessarily work equally well for both.
Looking for your own sector? The Other Sectors page points to the closest match.
Questions worth considering
- When a process is described as “too slow,” has anyone checked whether what is arriving at it has changed, not just how much of it arrives?
- If both volume and input quality are moving together, what would separate their individual effects — and is that worth doing before committing to a fix?
- What did the people who used this process well, when it worked, know that newer or more numerous users might not?
More on this topic: Growth.
Your team is facing similar challenges?
Subscribe to The Riddle Ledger
Perspectives are sent as they are published, with no promotional email, and you can unsubscribe at any time.