I recently remembered a strange experience on Uber’s infrastructure team, when I was asked by a senior engineer on the mobile development team to prioritize setting up and managing their mobile test clusters. At the time, most of my attention was focused on better defining the services we provided–eventually ending with a service cookbook–but as I reflect on it now, what comes to mind is something else entirely: how frequently teams and organizations make resourcing decisions that at least initially come across as entirely implausible.
How wonderfully bizarre that the mobile engineering team at Uber, a company whose product is exclusively accessible using a mobile app, a company closer to “mobile only” than “mobile first”, couldn’t get their mobile development needs prioritized for support.
The details of this example get a bit murkier as you dig in—we had some major scalability issues that were taking priority, there was confusion around who should be hiring for these roles, and eventually this was solved by hiring folks specialized in managing mobile test clusters-—but the murkiest thing of all is that this kind of prioritization perversion is a common pattern that repeats across companies, organizations and projects.
As my experience has extended, this pattern has trended towards universal, and I’ve become curious what inspires this pattern, and just how rational are the underlying decisions.
Why does this happen?
The most common cause of misunderstood priorities is that folks are operating on different timeframes. This concept is explored in depth within
The Innovator’s Dilemna,
and can be uncomfortably summarized as your company’s best product today not being your best product a decade from now, and your success depends
on investing into your pressing needs today while also planting seeds to sustain you through the future’s many tomorrow’s.
This dynamic creates an interesting tension between doing what’s best for your current business, products and organization and doing what’s necessary to facilitate the future’s genesis.
Even in cases where a company’s product is nominally the same, a decade later it’s typically several generations evolved. Continued success requires winning repeated rounds: success in early rounds supports but doesn’t guarantee future success.
The Uber example I started this piece with is an example of timeframe tension: the scalability fire required active investment, which was starving our ability to prioritize investing into infrastructure to support better mobile testing. On a longer timeframe, the decision to not invest into mobile test infrastructure was ludicrous, but in the short term making the investment was impossible.
Bandwidth to compete
Another source of confusing priorities is when gaining access to company resources becomes very time consuming, requiring numerous rounds of negotiation, with each round requiring a series of meetings, chats, presentations or proposals.
This is intended to ensure that the most important projects get effectively resourced, but an unintended outcome is that the neediest teams are simply unable to pay the participation tax to request additional resources. This leads to teams that are already well-staffed continuing to create the most compelling requests and to gain the most additional resources. This creates a veneer of optimal resourcing, but in practice only staffs the most impactful opportunities if you have maintained a consistently perfect track record of resourcing.
To prevent this scenario, reduce overhead for requesting resources, including capping the efforts that teams can inject into the process. Folks running resourcing processes should be extremely skeptical of teams or leaders who make repeated requests for resources, particularly in private and outside of the defined resourcing and planning regime.
It also helps for your leadership team to have a solid mental map of priorities before resourcing begins, and to remain skeptical of outcomes that understaff what ought to be most important.
Optionality versus indecision
There is a fine line between indecision and preserving optionality, with each often playing the wolf wearing the other’s sheepy disguise. Both can lead to difficult to understand resourcing decisions, typically taking the format of allocating resources across sufficient projects that none is well staffed.
This can be “peanut buttering”, which leads to every project partially complete and none delivering value, but sometimes it’s savvy to preserve optionality and invest broadly, even at times into parallel efforts towards similar goals. Typically this strategy makes sense if there will be early learnings that make it clear which approach to pursue (something along the lines of prototyping), or if accomplishing the projects’ goal is vital to company survival (some product adoption, cost efficiency and scalability efforts take this shape).
The essential bit here is that you should be honest about whether you’re doubling up resources because a project is truly critical, or if your halving resources such that projects won’t finish in a reasonable timeframe. The later is simply inefficient resourcing.
Some folks in some companies attain so much influence that their world view warps reality around them. This level of influence is almost always built on top of a chassis of success, where the folks who possess it have repeatedly done good work at the current or previous roles, which becomes the prime indicator of whether their current work or decisions are good.
Past success is not always a reliable predictor of future success, and this becomes most obvious in cases where someone needs to operate in a markedly different context than they’ve succeeded previously. This is very obvious when folks switch between companies, but can also sneak up on you when you’re in a rapidly growing company that has quietly changed underneath your feet without you noticing.
This one is particularly difficult, because influence is a lagging indicator, and new values are recomputated at a particularly sluggish rate, particularly at larger companies where folks only periodically work together closely. In such cases it can takes years for someone’s influence to reflect the value of their current work (positive and negative).
Is relying on influence irrational? Probably not! It’s an important short cut to allow faster decision making, particularly in environments where there is so much information that it’s expensive to fully load all the context into your head before making a decision. However, the thing to keep an eye out for when considering influence is that it’s a prime vector for bias. It’s easy to let our instincts latch onto aspects that are at best irrelevant, and when we do, relying on influence to factor into decision making soaks in irrationality.
In the end, I hope the takeaway is that we should look deeply into others’ decisions for reasons why they may be rational despite looking flawed from some vantage points, and likewise that we consider our own decisions for vantage points where they may be irrational. There are few decisions that don’t seem terrible from one perspective and astute from another, so we’re all best off by being careful in our decisions and in our judgement.