The three Ps of sustainability: what to measure beyond profit

A large customer sends suppliers a questionnaire, your bank asks about environmental impact, and you hear "the three Ps" more than once. What number goes against the second and third P? John Elkington, who says he coined "triple bottom line" in 1994, wrote in 2018 that thousands of such reports are produced each year, yet it is far from clear that the data help decision-takers (Elkington, 2018). This post is about that snag: measuring, and adding up.

What do the three Ps say, and what did Elkington want them to do?

An HBS Online blog post (Miller, 2020) describes the triple bottom line as a business concept holding that firms should commit to measuring their social and environmental impact, in addition to financial performance, rather than focusing solely on profit; the three Ps are profit, people and the planet (Miller, 2020). Its examples are general: fair hiring, partnerships with non-profits, ethically sourced materials, lower energy use, leaner shipping (Miller, 2020). The library's triple bottom line card carries the same kind of example.

Elkington writes differently. In his account, the original idea encouraged businesses to track and manage economic (not just financial), social and environmental value added or destroyed, and it fed into platforms such as the Global Reporting Initiative (GRI) and the Dow Jones Sustainability Indexes (DJSI). He writes that the framework was not designed to be just an accounting tool, yet many early adopters understood it as a balancing act with a trade-off mentality (Elkington, 2018). The sustainable marketing post already used this passage when discussing environmental claims; I come back to it here because it touches the measurement problem.

Why is it so hard to add the bottom lines into one number?

Norman and MacDonald split the framework into a few claims. The Measurement Claim says the components of "social performance" can be measured in relatively objective ways using standard indicators, then audited and reported. The Aggregation Claim says a social "bottom line", analogous to net profit or loss, can be calculated from those indicators using a relatively uncontroversial formula that could be used for any firm (Norman and MacDonald, 2004). In their view only the second claim is distinctive of a "bottom line" approach, and they argue it does not hold; as far as they know, no one has proposed a method for calculating it (Norman and MacDonald, 2004).

The first reason: financial accounting rests on a "unit of measure" assumption, that revenues, expenses and assets all reduce to a common unit of currency. On the social side there is no obvious common unit for all the good and bad a firm does to different stakeholder groups (Norman and MacDonald, 2004). They imagine a firm reporting that 20 per cent of its directors were women, 7 per cent of senior managers were from visible minorities, it donated 1.2 per cent of its profits to charity, hourly staff turnover was 4 per cent, and it was fined twice for toxic emissions. Out of context, they say, it is hard to know whether these figures are good or bad, and two percentages of different kinds, such as 20 per cent women directors and 7 per cent minority managers, cannot be added into a meaningful 27 per cent (Norman and MacDonald, 2004).

The second reason: the values cannot be weighed against each other. They offer three pairs: a generous family-friendly policy alongside three sexual-harassment suits in a year; an ethical sourcing policy audited by an international human-rights NGO alongside a bitter three-month strike at home; a donation of 2 per cent of gross profits alongside a price-fixing conviction. They ask whether there is any obvious way to say if each pair is a net gain or loss, and answer that weighing them takes value judgements on which reasonable people can disagree (Norman and MacDonald, 2004).

Does a long report make the numbers usable?

Elkington writes that the triple bottom line and its variants have been captured and diluted by accountants and reporting consultants. He notes thousands of reports a year, but says it is far from clear that the data are aggregated and analysed in ways that genuinely help decision-takers and policy-makers track, understand and manage systemic effects (Elkington, 2018). He also writes that CEOs and CFOs move heaven and earth to hit profit targets, yet the same is very rarely true of their people and planet targets (Elkington, 2018).

Norman and MacDonald show how a triple bottom line report can become hard to compare: the firm picks its own data, and may change its indicators over the years because it thinks the new ones are more relevant, or perhaps to thwart comparability. That makes one firm hard to follow over time, and comparison across firms often impossible. They sum the framework up as the old single bottom line plus vague commitments to social and environmental concerns (Norman and MacDonald, 2004).

What can replace a single total?

Neither source offers a usable table of indicators. What follows is my proposal, assembled from the snags they point to and from the test Elkington cites from Patrick Thomas, former chief executive of Covestro: progress on at least two dimensions while the third remains unaffected (Elkington, 2018). Four rules:

1. Each P gets one lead indicator, with its unit and method of calculation written down. Other indicators are secondary.
2. The three Ps are never added. Norman and MacDonald argue that this does not hold up.
3. Keep the indicator set for at least three years; if you change it, say why and give both calculations.
4. Apply Thomas's test as a reading condition: if one P gets worse, the report states that line first and does not bury it in a total.

This example is entirely hypothetical; "Company V" and every figure are mine and belong to no real business. Company V is an 80-person garment workshop.

Two Ps improved, so at a glance it looks fine. But the second P got worse, so under rule 4 the report opens with: staff turnover rose from 28 to 31 per cent, and this is what the board will do about it. Secondary indicators, such as lost-time accidents or the share of offcuts reused, sit below and do not offset the lead indicator. A "total score" is outside this proposal, because rule 2 rules out adding.

This suits you if you need a compact table to run the firm and to answer one specific customer or bank. It does not suit you if the asker already requires a particular reporting standard, since I have not read GRI or any other, or if you plan to use the table as the basis of a claim to consumers; for that, see the evidence section of the marketing post.

Going deeper: what did I read, and what could I not check?

I read the full text of Elkington (2018) in Harvard Business Review, taken from the article body on the magazine's page. I read Norman and MacDonald in full through the authors' manuscript (March 2003, marked "in press"); the library records the paper in Business Ethics Quarterly 14(2), pages 243-262, with different pagination, so I paraphrase and give no page numbers. I read Miller (2020) on HBS Online; it is a course-promotion post, offers no indicator guidance and cites no study for its claim on ESG returns. I have not read Elkington's book (1997), the origin recorded on the library card, nor GRI or the DJSI.

Three limits. First, Norman and MacDonald is a 2003 manuscript, so I do not know whether any aggregation method appeared later. Second, the four rules and the indicator table are my own proposal, untested on a real business; they combine Thomas's test (a second-hand report from Elkington) with the critiques above. Third, Company V and its figures are hypothetical.