In Isaac Asimov’s futuristic science-fiction short story “The Feeling of Power,” computers have taken so much control that humans have lost the ability to count let alone process math.
Almost 60 years after it was published, this dystopian tale doesn’t seem so far-fetched. One out of every five adults in the UK struggles with basic mental arithmetic, according to a BAE System survey conducted by YouGov. Of those surveyed, 35 percent will use a calculator to add numbers that exceed 100. More alarmingly, those ages 18 to 24 are almost three times as likely to use calculators for basic tasks compared with people 55 and older. Further, 13 percent of those 55 and older make mistakes while 22 percent of those younger than 45 make mistakes when performing random simple multiplication tasks (especially 11 times 11). The U.K. seems to be representative of most Western countries. U.S. adults appear to score lower.
There’s little margin for error in the business world. Managers who influence strategy and decisions must see through the numbers to prevent foolish and expensive mistakes. F1F9, a U.K. based agency that helps blue-chip companies conduct financial modelling, highlighted how common and devastating spreadsheet mistakes can be in its report “Capitalism’s Dirty Secret.” Mistakes are rarely documented. When they are – such as JPMorgan’s $6 billion London Whale loss – they are fairly hard to believe.
Paradoxically, as we move from calculators and excel sheets to machine learning and artificial intelligence, the ability to make quick sense of numbers and ensure they are reasonable is not just more rare but more required.
Companies today seek more granular analysis that creates massive data cubes across business lines, geographies and channels all the way down to customer-level contribution margins. Prediction models increasingly aim to forecast profit opportunities that will better guide resource allocation and operational priorities. Yet it is anachronistic to rely on large teams of analysts to manage all these data because advanced algorithms can massively improve prediction accuracy and new digital tools enable unprecedented targeting with precision and personalization.
However, strategic fundamentals have not changed. A business still needs to get its customers to pay more than what it costs to serve them. Most business cases have just a handful of assumptions that REALLY mater. Is average selling price realistic relative to the volume and market share it represents? Are costs consistent with benchmarks and past experience and is the comparison made on a like-for-like basis? Is the cash profile manageable to sustain the investment and cover contingencies?
Senior managers must rapidly determine which numbers make sense so they can identify errors and prevent catastrophic mistakes. Properly reviewing plans and pressure testing assumptions almost always requires real-time recalculation of pivotal numbers.
We famously emphasize this skill in our recruiting interviews, but, more importantly, we constantly apply it at work. Much of this problem-solving approach begins with “back of the envelope” calculations that many of us actually do with no paper at all. Hypotheses are made about impact and contributing factors, adding up the assumptions to estimate the value and the break-even points. As teams launch into intensive analytics, they start to compare the findings with the initial hypotheses. If there is a discrepancy, they study whether there are errors in the model or if the initial assumptions were wrong. I will not bring an important number (e.g. of estimated impact) to a meeting with senior clients if I cannot explain and multiply key assumptions without a calculator. As I prepare, I sometimes find crucial errors even at rehearsal stages.
Much greater math fluency is mandatory to overcome the confirmation bias that often results in the initial estimation. I was working with a private-equity client a long while back who was looking to purchase a utility that was being privatized. A fairly complicated regulatory pricing model was in place, and I just couldn’t make the formula work! My calculation was many times different than the regulator’s examples and completely undermined the PE firm’s business case. Despite my self-doubt, I kept plugging away until I was certain the formula was in error. I convinced the clients of the problem and drafted a letter for them to send to the regulator explaining what was wrong. The pushback delayed the auction by a full year, but the formula was corrected.
In a post a few months back I described that too many predictions “don’t add up”, i.e. sometimes a very sound set of assumptions may make sense in isolation but not when you add up multiple business cases and realize that the aggregated result may be outlandish. Some of the best CEOs I have seen (benefiting from the bird’s-eye view) continuously use mental math to compare between initiatives and correct for that.
Mental math won’t only help improve your business-decision making ability, it will help you save time (if you estimate the activity’s duration better) and money (when a shopping or restaurant bill is miscalculated). Perhaps, if you practice your proficiency with probabilities, you can better understand how the world works, not to mention see an increased ability to determine whether you need an umbrella or pair of sunglasses. Despite all my love for math, I don’t get that right often!
Yuval Atsmon is a senior partner in McKinsey’s London office.
Originally published on LinkedIn.