Performance Measurement Systems & the Balanced Scorecard

Financial Performance & Management Control - TUHH Institute of Management Accounting & Simulation, Hamburg · part of my Technology Management MBA · study notes for revision.

This is the final chapter of the course. Chapter 11 looked at single performance measures: what they are, the four ways to type them (local or global, leading or lagging, relative or absolute, financial or non-financial) and the control problem that appears the moment you pay people on them. This chapter takes the next step: nobody steers a company on one number, so how do you assemble many measures into something that hangs together?

Two answers dominate. The DuPont system is the old, purely financial, purely arithmetic answer: take ROI and pull it apart into the pieces that drive it. The Balanced Scorecard (BSC) is the modern answer: keep the financial numbers, but put them next to customer, process and learning measures, and connect all four through a chain of cause and effect that mirrors the strategy.

The chapter answers four questions. What types of performance measurement systems exist? How does the BSC go beyond the traditional systems? What does a real BSC look like? And what is the design logic behind one? The Johansen department-store case then shows what goes wrong when a company pays its managers on three financial numbers and nothing else.

1 · From single measures to measurement systems

Chapter 11 defined a performance measure as quantitative information that compresses a complex reality into a number, so that a manager can be informed quickly. The catch is in the word compresses. ROI on its own tells me a division earned 12% on its capital, but not whether that came from fat margins on slow sales or thin margins on fast sales, whether it is rising or falling, or whether it was bought by starving the store of staff. To make single measures more meaningful, they are combined into performance measurement systems: sets of measures that are arranged so that each one gives context to the others.

That is the whole purpose of a system, and it gives us the two questions to ask of any system we meet: how are its measures connected, and what mix of things do they look at?

2 · Classifying measurement systems - linkage and balance

The course uses a simple two-by-two scheme with two dimensions:

Connection (linkage) of performance measures - are the measures tied to each other, arithmetically or through explicit cause-effect logic, so that a change in one can be traced to the others? Or are they just a loose list?
Balance of performance measures - do the measures cover different kinds of things (financial and non-financial, results and their drivers, inside and outside the firm), or do they all look at one dimension?

Selective performance measures low linkage · low balance

A handful of stand-alone numbers picked case by case
No formal connection between them, usually all financial
The starting point most firms never leave

DuPont system high linkage · low balance

ROI split arithmetically into margin, turnover and their drivers
Every number is mathematically tied to the top
But every number is financial and backward-looking

EFQM model lower linkage · high balance

The European quality-management framework: many criteria, wide coverage
Very balanced across people, processes, society and results
The criteria are not tightly linked to one another

Balanced Scorecard, value-driver hierarchies high linkage · high balance

Financial plus non-financial measures, deliberately balanced
Connected by cause-effect chains (BSC) or driver trees (value-driver hierarchies)
The target quadrant of modern performance management

Read the grid as a map. Bottom-left is where most firms start: a few selective numbers. Moving right adds linkage (DuPont). Moving up adds balance (EFQM). The top-right corner has both, and that is where the Balanced Scorecard and the value-driver hierarchies sit.

3 · The DuPont system - perhaps the oldest system in the world

The course calls the DuPont System of Financial Control maybe the oldest performance measurement system in the world; the version shown traces back to a description from 1950. Its idea is beautifully simple. Take the top measure a division is judged on, return on investment, and split it into two factors that can be managed separately:

ROIROI = profit margin × capital turnover

Profit margin margin = profit ÷ sales - how much of each sales euro survives as profit

Capital turnover turnover = sales ÷ invested capital - how many euros of sales each euro of capital generates per year

Multiply the two and sales cancels, leaving profit ÷ invested capital, which is ROI. Each factor is then split again into the things that actually move it:

Return on investmentprofit ÷ invested capital - the top of the tree

↓= profit margin × capital turnover

Profit marginprofit ÷ sales

Profit= revenues minus costs

Driversrevenues on one side, the cost blocks on the other

Capital turnoversales ÷ invested capital

Invested capital= fixed capital + working capital

Driversplant and equipment; inventories, receivables and cash

The DuPont tree. ROI at the top, the two factors below it, and under each factor the balance-sheet and income-statement items that drive it. Every branch is connected to the root by pure arithmetic, so a change anywhere can be traced up to its effect on ROI.

Why is it high linkage, low balance? High linkage because the whole tree is one identity: change inventories and you can compute exactly what happens to turnover and therefore to ROI. Low balance because every node is a financial figure from the statements - there is nothing about customers, quality, staff or the future. It is a superb tool for explaining a financial result, and a poor tool for predicting the next one.

4 · Which dimension matters more? Balance beats linkage

If a firm can only improve one of the two dimensions, which should it pick? A study cited in the course (2003) modelled how satisfied managers are with their performance measures as a function of the two dimensions. The path coefficients were roughly 0.53 for balance and 0.22 for linkage, both significant at the 1% level, and together they explained about 44% of the variance in satisfaction.

Balance of measurespath coefficient about 0.53

→strong effect

Satisfaction with the measuresexplained variance about 44%

←weaker effect

Linkage of measurespath coefficient about 0.22

Both dimensions help, but balance helps more than twice as much as linkage. Managers are happier with a broad set of loosely coupled measures than with a tightly linked but one-sided one.

The practical reading: a beautifully linked DuPont tree still leaves managers unsatisfied, because it only answers what happened to the money. Adding non-financial, forward-looking measures does more for the usefulness of the system than adding more arithmetic. That is the empirical case for the Balanced Scorecard.

5 · The Balanced Scorecard - four perspectives

The Balanced Scorecard of Kaplan and Norton is the best-known modern performance measurement system. It keeps the financial measures but complements them with three further perspectives. The course stresses that the four classic perspectives are not a law of nature: they are a suggestion based on empirical experience. Two observations lie behind it:

Successful companies set themselves goals in every one of these areas in order to excel - not just financial targets.
An unsatisfactory result in a single perspective is enough to endanger the success of the whole company for a long time. You cannot compensate neglected staff or angry customers with one good quarter.

Financial perspective how do we look to our owners?

Profit, growth, returns, cash flow
The lagging outcome of everything else

Customer or market perspective how do customers see us?

Satisfaction, loyalty, complaints, market share
Where the strategy meets the outside world

Internal process perspective what must we excel at?

Quality, speed, availability, cost of the key processes
The processes that create the customer value

Learning and development perspective can we keep improving?

Skills, motivation, health, culture, information systems
The foundation the other three stand on

“Balanced” means balance in several directions at once: financial and non-financial, lagging results and leading drivers, the outside view (owners, customers) and the inside view (processes, people), short term and long term.

Goals, measures, targets, initiatives

A perspective is not just a list of numbers. For each perspective the design fills four columns, and it is this structure that turns a scorecard into a management tool rather than a report:

Column	Question it answers	Department-store flavour
Goal (objective)	What do we want to achieve here, given the strategy?	”Be the service leader in every store”
Measure	Which number tells us whether we are getting there?	Customer satisfaction score from the survey
Target	What value counts as success, by when?	Average of at least 4.4 out of 5 this year
Initiative	Which concrete actions will move the measure?	Service training, more floor staff at peak hours

6 · The cause-and-effect chain (strategy map)

The measures of a BSC must not simply be listed in four boxes; they have to be connected in a cause-effect way. The chain always runs in the same direction, and that direction is the argument for the whole tool:

Learning and developmentskilled, motivated, healthy people with the right tools

↓enable

Internal processesfaster, better, more reliable key processes

↓deliver

Customer and marketsatisfied, loyal customers and a growing share

↓produce

Financial resultssales growth, margin, return, cash flow

The strategy map. Each layer is a hypothesis about the one below it: if we develop these capabilities, then these processes improve, then customers respond, then the financial numbers follow. The financial perspective sits at the end of the chain, not at the beginning.

This is what “the BSC goes beyond traditional systems” means. DuPont links measures arithmetically and only within finance. The strategy map links measures causally and across perspectives: the non-financial measures are the leading indicators of the financial ones, so the scorecard can warn you before the money moves. It also makes the strategy testable: if training hours go up and process quality does not, the hypothesis was wrong and the map needs revising.

The design logic of a BSC

Translate the strategy into goals - two to four objectives per perspective that say what winning looks like there.
Draw the cause-effect chain - state explicitly which learning goals drive which process goals, which customer goals, and finally which financial goals.
Choose one or two measures per goal - a mix of lagging results and leading drivers, financial and non-financial.
Set targets - the value and the deadline that count as success, adjusted for what the unit can actually influence.
Fund initiatives - the concrete projects that should move the measures, and the budget behind them.
Review the hypotheses - when a leading measure improves and the lagging one does not, fix the map, not just the number.

7 · A real example - the German railway’s headquarters scorecard

The course shows the headquarters version of Deutsche Bahn’s scorecard (its “BahnStrategieCard”). It uses four fields that map neatly onto the classic perspectives, even though the labels are the railway’s own:

Railway field	Classic perspective	Measures on the card
Efficiency / financial goals	Financial	Sales; productivity; operating profit and contribution margin; net cash flow
Customer satisfaction / market share	Customer and market	Result of the customer survey; customer loyalty; complaints; development of the specific market share
Quality of services rendered	Internal process	Punctuality; availability of vehicles, facilities and personnel; customer relevance and service quality
Commitment of employees	Learning and development	Health status; performance orientation and management culture; improvements; employee qualification

Notice the cause-effect logic hiding in the table. Healthy, qualified, improvement-minded staff (bottom layer) make trains punctual and available (process layer), which shows up in survey results, loyalty and fewer complaints (customer layer), which finally feeds sales, productivity and cash flow (financial layer). The railway simply renamed the four boxes to fit its own language.

8 · Actual usage - well known, rarely fully used

The BSC is one of the most famous management tools of the last decades, yet a survey of the 200 largest listed German firms found that only about 24% really use a Balanced Scorecard. Many others run partial versions, for example a scorecard that has the four boxes but no cause-effect chain, and some run purely financial “scorecards” that carry the name but none of the balance.

Worked example

(a) A DuPont tree with real numbers

Take company 3 from the financial statement analysis chapter: sales 7,123, operating income 1,832, total assets 14,630. Using operating income over total assets gives a return on assets, but the DuPont logic is identical to ROI.

Margin1,832 ÷ 7,123 = 0.257 = 25.7%

Turnover7,123 ÷ 14,630 = 0.487 ≈ 0.49

ROA 25.7% × 0.487 = 12.5% - check: 1,832 ÷ 14,630 = 12.5%

Return on assets 12.5%operating income 1,832 ÷ total assets 14,630

↓split into

Profit margin 25.7%operating income 1,832 ÷ sales 7,123

↓multiplied by

Capital turnover 0.49sales 7,123 ÷ total assets 14,630

↓built from

The raw figuressales 7,123 · operating income 1,832 · assets 14,630

Company 3 through the DuPont lens: a very fat margin (a quarter of every sales euro is operating profit) sitting on a slow capital turnover (each euro of assets produces only about half a euro of sales a year). The 12.5% return is a margin story, not a volume story.

What the tree tells a manager that the bare 12.5% does not: the lever with room to move is turnover. Every 0.1 of extra turnover at the same margin would lift ROA by about 2.6 percentage points, so the questions go to inventories, receivables and idle fixed assets, not to prices.

(b) A mini Balanced Scorecard for a department store

Learning and developmenttraining hours per associate · staff turnover · internal promotions

↓trained, stable staff run better floors

Internal processesstock availability of advertised items · checkout waiting time · returns handled on the spot

↓better floors create better visits

Customersatisfaction score · repeat-visit rate · complaints per 1,000 transactions

↓loyal customers spend more, more often

Financialsame-store sales growth · gross margin · store operating profit

A store scorecard with two or three measures per perspective and the strategy map that connects them. Read upward from the bottom to see why the financial box is a result, read downward from the top to see what has to be true first.

The hypotheses in words: if turnover falls and training rises, then stock is on the shelf and queues are short; if that holds, then satisfaction and repeat visits rise and complaints fall; if that holds, then same-store sales grow at a healthy margin. Each arrow is a claim that can be checked against the data a year later.

Case corner - Johansen’s New Scorecard System

The set-up. Johansen’s is a large high-end department-store chain, founded in New York in 1950 and built on one principle: superior customer service. By 2014 it had 121 stores in 32 states, organised in five regions, each with a regional manager over the store managers; the firm liked to promote from within.

The financial-only system. In 2006 the company hit financial trouble, most senior executives were replaced, and the new team introduced a financially based incentive system. Store managers had real autonomy over sales levels, the in-store experience, local marketing and promotions, merchandising and staff training, so the stores were treated as profit centres and managers were rated and paid on three financial metrics: same-store sales growth, gross margin and net income. What they did not control: store location (and therefore customer demographics) and other major investments.

What it did to behaviour. The three numbers rewarded exactly one thing: money this year. Cost cutting lifted profit, and the study later found that those same cuts had fuelled compromises in customer service. Nobody had told managers that service still counted; the leadership had simply assumed they knew. Incentives drive behaviour, and the incentive said “cut”.

Why performance stalled. After a modest early improvement, results had stagnated again by 2012: net sales grew from about 6,725 in 2006 to about 7,790 in 2012, but net income as a share of sales stayed stuck between roughly 5% and 7% the whole time. A commissioned study gave three reasons: e-commerce had been the fastest-growing retail sector since 2008 and was taking a rising share of spending, and Johansen’s could not win a price war against online cost structures; market research showed a rival’s customer service had become superior; and the data suggested the profit-driven cost cutting had eroded the very thing that differentiated the chain. Nearly all sales still came from the in-store experience, so losing service leadership was existential.

The new scorecard (from 1 January 2013). Store managers are now rated 1 to 5 in four categories:

Financial - year-on-year sales growth and profitability against a baseline target (needed for a 3) and a stretch target (needed for a 5), with baselines set by the regional manager and corporate finance and adjusted for each store’s demographics.
Customer service - a third-party survey (10 questions, each scored 1 to 5, averaged to a store score); a response rate of at least 12% is required for a 4 or 5, a score of at least 4.4 for a 5, plus a qualitative reading of the comments. Company-wide in 2012 the response rate was 18.5% and the average score 3.6.
Leadership - annual 360-degree feedback, staff turnover and employee complaints, weighed by the regional manager.
Strategy - promotion of the brand and branded merchandise, and implementation of corporate initiatives such as inventory management, training and merchandising mix.

The overall rating is not an average. It is a judgement by the regional manager under corporate rules: a 5 overall needs at least a 3 everywhere, a 5 in three of the four categories, and at least a 4 in customer service. Bonus is 10%, 25% or 40% of salary for an overall 3, 4 or 5, and the ratings are debated at an annual performance summit so they are consistent across regions; the COO steps in only when the regional manager and the group disagree.

Store 51 and the open issues. The flagship in Orange County (revenue 150 million, gross margin 40% against a company average of about 35%, affluent older customers, no direct competitor nearby) beat every financial stretch target, yet the new lens exposed 45% staff turnover in the first half-year and satisfaction scores of 3.2 and 3.4. The manager improved both (turnover down to 33%, below the 35% company average; satisfaction up to 4.1 by the fourth quarter), but the survey response rate never rose above 7%, so the customer-service rating was capped at 3 and the rules give an overall 4, while the regional manager argues for a 5. The dilemma bundles every open issue of scorecard design:

Weighting - the rules make customer service the gate to the top rating; is that the right weight, and should the overall rating be a formula or a judgement?
Controllability - the response rate depends on demographics (older, wealthier customers go online less), which the manager did not choose; baseline targets are adjusted for demographics, the survey threshold is not.
Gaming - associates highlight the survey instructions on receipts, upset customers are more likely to answer, and under the old system cost cuts could buy a bonus; every measure invites its own manipulation.
Too many measures and too much subjectivity - four categories, sub-rules and qualitative judgements make the rating richer but also harder to compare across regions and easier to argue about.
Buy-in - financially minded managers hired during the profit push feel exposed to “soft” metrics.

What the case teaches. Pay on three financial numbers and you get three financial numbers, at the expense of the strategy. A balanced system rates managers on what they can influence (service, staff, execution), places the financial result at the end of the causal chain, and hard-wires the differentiator into the rules. And even a good design leaves judgement calls that no formula removes.

Key terms

Term	What it means in plain words
Performance measurement system	A set of measures arranged so they give each other context; built because single numbers alone are not meaningful enough
Linkage (connection) of measures	How tightly the measures are tied together, arithmetically (DuPont) or causally (BSC)
Balance of measures	How broadly the measures cover different dimensions: financial and non-financial, lagging and leading, inside and outside
Selective performance measures	A loose handful of stand-alone numbers; low linkage, low balance
DuPont system	The oldest system: ROI decomposed into profit margin times capital turnover and their drivers; high linkage, low balance
Profit margin	Profit divided by sales; how much of each sales euro is kept
Capital turnover	Sales divided by invested capital; how hard the capital works
Value-driver hierarchy	A tree that breaks a top value measure down into operational drivers; high linkage and, if non-financial drivers are included, high balance
EFQM model	The European quality framework with many criteria; high balance, weaker linkage
Balanced Scorecard	A system with financial, customer, process and learning perspectives linked by cause-effect chains; goals, measures, targets and initiatives per perspective
Strategy map	The cause-effect chain of a BSC running from learning through processes and customers to financial results
Leading vs lagging measure	A leading measure moves first and predicts (training hours, satisfaction); a lagging measure records the outcome (profit, ROI)
Profit centre	A unit whose manager is judged on its profit, as Johansen’s stores were from 2006
Controllability principle	Judge managers only on what they can influence; Johansen’s store managers cannot choose location or big investments
Gaming	Improving the measure without improving the thing it stands for, e.g. cost cuts that lift margin but hurt service

Test yourself

Why are single performance measures combined into systems, and which two dimensions does the course use to classify those systems? Place selective measures, the DuPont system, the Balanced Scorecard and the EFQM model on the grid.
A company reports sales of 7,123, operating income of 1,832 and total assets of 14,630. Compute the profit margin, the capital turnover and the return on assets, and say which of the two factors has more room to improve.
What did the research on managers’ satisfaction with their measures find about balance versus linkage, and what does that imply for a firm that runs only a DuPont tree?
Name the four perspectives of the Balanced Scorecard and describe the direction of the cause-effect chain. Why does a poor result in a single perspective matter so much?
Map the four fields of the German railway’s headquarters scorecard onto the classic perspectives and name two measures in each.
Johansen’s paid store managers on same-store sales growth, gross margin and net income. Explain how this eroded the strategy, what the new scorecard added, and two design issues that remain open.

A single measure compresses reality so far that it loses meaning; a system restores it by giving each measure context. The two dimensions are connection (linkage) and balance. Selective measures: low and low. DuPont: high linkage, low balance. BSC and value-driver hierarchies: high and high. EFQM: high balance, lower linkage.
Margin = 1,832 ÷ 7,123 = 25.7%. Turnover = 7,123 ÷ 14,630 = 0.49. ROA = 25.7% × 0.49 = 12.5% (check: 1,832 ÷ 14,630). The margin is already very high; the slow turnover is the lever, so look at assets, inventories and receivables.
Balance had a path coefficient of about 0.53 and linkage about 0.22, both significant at 1%, explaining about 44% of the variance in satisfaction. A DuPont-only firm has maximised the weaker dimension; adding non-financial, leading measures would do more for usefulness than more arithmetic.
Financial, customer or market, internal process, learning and development. The chain runs learning → processes → customer → financial. The perspectives are based on the empirical observation that successful firms set goals in all four, and that failure in even one is enough to endanger the whole company for a long time.
Efficiency/financial = financial (sales, productivity, operating profit and contribution margin, net cash flow). Customer satisfaction/market share = customer (survey results, loyalty, complaints, specific market share). Quality of services = process (punctuality, availability of vehicles, facilities and staff, service quality). Commitment of employees = learning (health, performance orientation and management culture, improvements, qualification).
The three metrics rewarded short-term money, so cost cutting lifted profit while customer service, the firm’s differentiator against e-commerce and a better-serving rival, slipped; results stagnated by 2012. The new scorecard rates managers 1 to 5 on financial, customer service, leadership and strategy, requires at least a 4 in customer service for a top rating, and ties bonus to the overall rating. Open issues: weighting and the subjective overall rating; controllability (survey response rates depend on demographics); gaming of surveys and cost cuts; the number and subjectivity of the measures; buy-in from financially minded managers.

Revision summary

The one-minute recap

Single measures compress too much; performance measurement systems combine them to make them meaningful.
Classify systems by linkage and balance: selective measures (low/low), DuPont (high linkage, low balance), BSC and value-driver hierarchies (high/high), EFQM (high balance, lower linkage).
DuPont: ROI = profit margin × capital turnover; margin = profit ÷ sales, turnover = sales ÷ invested capital; below that, revenues and costs, fixed and working capital. Pure arithmetic, purely financial.
Research: balance (about 0.53) beats linkage (about 0.22) in explaining managers’ satisfaction with their measures (about 44% of variance).
BSC: four empirically grounded perspectives - financial, customer/market, internal process, learning/development; successful firms set goals in all four, and a single weak perspective can endanger the firm for a long time. Each perspective carries goals, measures, targets, initiatives.
The perspectives are linked by a cause-effect chain (strategy map): learning → processes → customer → financial. Non-financial measures lead, financial measures lag.
Example: Deutsche Bahn’s headquarters card with efficiency, customer, service-quality and employee-commitment fields. Reality check: only about 24% of the 200 largest listed German firms really use a BSC.
Johansen’s: three financial metrics bought short-term profit and lost customer service; the new four-category scorecard restores balance and gates the top rating on service, but weighting, controllability, gaming and subjectivity stay open.

Back to the course overview →.