The Defensive Dashboard

In the early 2000s three financial economists, John Graham, Campbell Harvey and Shivaram Rajgopal, put a question to 401 chief financial officers that really allows only one answer: would you start a project that demonstrably brings the company far more than it costs? 55 percent replied that they would pass on it if the project pushed them below analysts' earnings expectations in the current quarter. Almost 80 percent said they would cut spending on research, advertising and maintenance in order to hit an earnings target. well documented What is remarkable about these answers is less the short-sightedness one can read into them than the calm with which they were put on the record, because in the world where these people are measured, they are simply being reasonable.

Every company tells its employees through metrics which behavior was just right, and it does so daily and at every level, from the board's quarterly figure to a project team's velocity. Metrics thereby form the signal layer of a company's action infrastructure, and like any infrastructure they act most strongly where no one is looking anymore. The previous article on the two gravities described why a company would rather avoid effort than improve the result; this one shows the place where that gravity does its work every single morning.

01Analysis: What is on the screens

Whoever leafs through the dashboards of an average mid-sized company or corporation finds strikingly similar numbers at every level. Read not as measuring instruments but as instructions, they yield a fairly precise picture of what the organization actually teaches its people.

The pattern runs through all eight rows, and it has a cause that lies deeper than the timidity of individual managers. The metrics themselves are built so that they reward defense faster, more visibly and more reliably than any advance. Five structural properties explain this.

Time asymmetry. A new venture costs money immediately, certainly and in full, while its return arrives later, uncertainly and often spread across units that do not even recognize it as such. Accounting sharpens this tilt, because research costs may be capitalized neither under German commercial law (HGB § 248 (2)) nor under IAS 38, so that in the books every innovation first appears as a deterioration, often over several years. The survey cited at the outset shows how executives respond to this. well documented

The denominator problem. Most steering figures are ratios, that is margin, return on capital or cost per unit, and a ratio is improved most comfortably by shrinking the denominator, investing less, outsourcing or discontinuing something. Clayton Christensen described this in 2008 with Stephen Kaufman and Willy Shih in „Innovation Killers“. well documented Boeing supplied the textbook case: return on net assets drove the far-reaching outsourcing on the 787, even though its own engineer L. J. Hart-Smith had calculated in an internal 2001 paper that the costs of outsourcing were being systematically underestimated. case

Invisible omission. The project that was never started has no line on any dashboard, and so the person who tries something and fails shows up red in the table, while the person who tries nothing appears nowhere. Psychology has known the tendency to weigh harm through action more heavily than harm through inaction since Ritov and Baron 1990 as the omission bias. well documented The metrics system amplifies this tendency, because it cannot technically represent omission at all.

Variance logic. Controlling almost everywhere compares plan and actual, and thereby rewards not the value created but the accuracy of one's own forecast. The consequences are well known: cautiously set plans, the December fever in which leftover budgets are spent so they will not be missing the following year, and a quiet aversion to anything that scatters. Michael Jensen took this apart in 2001 in „Corporate Budgeting Is Broken, Let's Fix It“. well documented Innovation is by its nature scatter, which is why a system that punishes scatter reliably prevents it.

Precision beats relevance. A cost figure is exact, attributable to one person and available weekly, while the value of an option on a future business stays vague, belongs to everyone and shows up late. In the meeting the precise thing wins the attention, and whoever has the attention sets the agenda. Daniel Yankelovich named this mechanism after the US secretary of defense Robert McNamara, who steered the Vietnam War by countable quantities.

02Hypothesis: Behavior follows how the number is built

From these five properties follows a hypothesis that can be tested against any company: whoever knows how a metric is built can largely predict the behavior of the people measured by it, regardless of what the mission statement says. The prediction rests on six diagnostic questions that can be put to any single metric.

The hypothesis does not claim that people react cynically to numbers, but that they learn, and a metrics system is a teacher that teaches every day and whose curriculum no one wrote. hypothesis In its building blocks it is well supported; as a whole-organization model it has not yet been systematically tested.

03Problem: Why this gets expensive

Costs can be lowered to zero at most, while revenue has no ceiling, and in this plain asymmetry lies the real problem of a defensive metrics system: it steers the entire attention of an organization onto the side of the ledger whose return is bounded, and pulls it away from the side whose return is not.

3M demonstrated this shift. For decades the company had set itself the goal of earning around 30 percent of its revenue with products only a few years old. After the introduction of Six Sigma under James McNerney from 2001, that share fell to around 21 percent by his departure in 2005, and the magazine BusinessWeek described in detail in 2007 how efficiency standards had smothered research. case At Boeing, the then head of the commercial airplanes division, Jim Albaugh, said in 2011 of the 787 program: „We spent a lot more money in trying to recover than we ever would have spent if we'd tried to keep the key technologies closer to home.“ case

The damage arises not in a single moment that would show up on a dashboard, but as the sum of thousands of small decisions, each reasonable on its own, until an organization has unlearned its capacity for the new and only notices when a competitor shows it up. The workforce probably changes along the way, too, because people who like to dare something are seldom promoted in such a system and leave earlier. hypothesis

A metrics system is a teacher that teaches every day and whose curriculum no one wrote.

04Approach: Redirect the loss aversion

Whoever designs a dashboard is laying track on which a thousand small decisions will later roll of their own accord, and the most effective rebuilds do not try to overcome people's loss aversion but to set it onto a different track. The precondition for all the levers below is the coupling from the core thesis: display, payout and promotion must point the same way. The letters of the levers reappear in the dashboard example below.

05Example: The same division dashboard, built twice

The following example shows a fictional business division with an identical situation, once in the usual build and once rebuilt along levers A to G. All figures are example values. Both dashboards show the same company, but they put different questions to its leadership.

06Goal and impact

The goal of the rebuild is a company whose employees learn from their metrics every day that calculated bets, quick decisions about whether to continue them, and the reallocation of savings into growth are the right behavior, without the discipline in the core business being lost.

The expected effects can be named and measured. Savings are sought as a source of funding rather than as an end in themselves, which shows up in the reallocation rate. hypothesis Ventures are started earlier and ended earlier, which shows up in shorter experiment cycles and a kill rate well above zero. hypothesis Planning becomes more realistic, because corridors offer less incentive for cautious setting than pinpoint plans, as the experience of beyond-budgeting companies suggests. case In the long run the vitality index rises, and it is the metric against which the whole rebuild must in the end be measured.

The risks can be named just as well. Innovation metrics are easier to game than cost metrics, because a figure like „number of ideas“ produces only idea theater without a counter-metric, which is why every future metric needs a kill rate or a documented stopping criterion beside it. A result before future investments invites relabeling operating costs, which can only be prevented by clear criteria for a bet and by their regular review. And the rebuild does not work as long as bonuses and promotions still hang on the old figures, because people read the payout more closely than any dashboard.

A company does not first have to learn to love risk for this; above all it has to stop reminding itself every morning first of what it might lose. Which picture it should see first each morning instead is one of the most consequential leadership decisions there is, and it is usually made by someone who believes they are only building a report.