When SDGs truly travel with strategy: five conditions to check

A European buyer's purchasing team sends your factory a questionnaire: "Which SDGs does your company align with? Please attach evidence." The deadline is two weeks. This scenario is my own invention, and your factory of a few hundred people has no sustainability department. What do you fill in, and what do you decline to fill in?

What does conditional SDG alignment mean?

It means testing a commitment before accepting it, rather than accepting first and hunting for proof later. The library's framework asks: when a leadership team weighs committing to the Sustainable Development Goals, how can it tell whether the commitment truly strengthens strategy or is just a label or a budget line? Conditional SDG alignment has five nodes.

The library's "origin" field lists five works and credits no individual author with the framework. The entry holds only the five node names, so the explanations below are my reading.

Why test at all? Van Zanten and Van Tulder (2021) draw, in their introduction, on surveys showing that the SDGs tend to sit with CSR or communications departments and that many companies pick easy, isolated goals to report on (Van Zanten and Van Tulder, 2021). Montiel et al. (2021) offer a reason: the SDGs are country-level targets for governments, and the SDG Compass leaves firms to interpret how to build them into operations (Montiel et al., 2021). When a buyer asks "which SDGs", the answer you write has to survive checking.

Condition 1: what must never be traded off?

Ethical red lines are the things that cannot be exchanged for SDG credit. The SDG Compass sets baseline responsibilities for every company: comply with the law, uphold internationally recognised minimum standards, respect universal rights. On human rights it says: "This responsibility cannot be offset by any effort to promote human rights or advance sustainable development" (GRI et al., 2015, p. 10).

Carroll's CSR pyramid (1991) has four layers according to the library: economic, legal, ethical, philanthropic. I could not open the original. Van Zanten and Van Tulder report, in the same introduction, that Carroll noted profit is among a company's most important responsibilities because all the others depend on it (Van Zanten and Van Tulder, 2021). My reading is that a voluntary SDG programme sits on top of compliance and cannot stand in for it. A factory with open safety findings should not lead its reply with a tree-planting scheme.

Both climate documents put emission cuts ahead of offsetting (see the comparison below). That is a boundary for carbon credits, not for labour rights, which I did not assess.

Condition 2: do money, people and contracts move with the goal?

Adaptive resource allocation means the goal has a budget, an owner and room to change when conditions change. My own question: when revenue falls, what is cut first? The pages I read do not answer it.

Unilever says its climate action areas are built into each Business Group's financial plan (Unilever, 2024, p. 13) and that it co-funds expert support for suppliers (Unilever, 2024, p. 19). Apple uses a different instrument: its Supplier Code of Conduct now requires the whole direct manufacturing supply chain to use 100 per cent renewable electricity for Apple production before 2030 (Apple Inc., 2025, p. 25). Suppliers relied mostly on renewable energy certificates, which Apple calls, for well-designed programmes, an important interim solution (Apple Inc., 2025, p. 26). Ask your buyer whether this is a requirement or support, and who carries the cost.

Condition 3: where are results measured, against what baseline, and who checks?

Outcome evidence is a figure with a base year, a scope and a verifier. Apple reports gross emissions down by more than 60 per cent from 2015, not including offsets, while revenue grew by more than 65 per cent (Apple Inc., 2025, p. 4). A different kind of number is the 41 million tonnes "avoided" in 2024, which a footnote says can exceed the reduction from the baseline (Apple Inc., 2025, p. 12). Do not read the two as one. Apex Companies assured selected data (Apple Inc., 2025, p. 104). Fraunhofer IZM reviewed the product-related Scope 3 footprint remotely, and Apple withdrew its first reported results on 28 February 2025 to revise how it counts component yield, so figures rose slightly (Apple Inc., 2025, p. 106).

Unilever's record differs. Absolute emissions from operations are down 74 per cent on 2015, but "achieving significant absolute reductions in our Scope 3 emissions has proven more challenging" (Unilever, 2024, p. 5). Actions cover about two-thirds of near-term Scope 3 targets, "with one-third left to find in the coming years" (Unilever, 2024, p. 13). Unless stated otherwise, the data "has not been audited or assured" (Unilever, 2024, p. 51).

One limit applies to both. Van Zanten and Van Tulder's framework (section 4.3.3) assesses a company's influence on social-ecological outcomes, not ultimate change in them (Van Zanten and Van Tulder, 2021). Your emissions falling does not prove an SDG moved.

Condition 4: who bears the cost, and who is asked?

Stakeholder legitimacy asks whether those affected accept how the standard is set, above all when the buyer sets it. Banerjee (2008) argues that CSR and sustainability discourses are defined by narrow business interests and serve to curtail the interests of external stakeholders (Banerjee, 2008). I use that as a question, not as a verdict on any company.

Unilever defines supplier "climate leadership" itself: science-based targets, public progress reporting, and the ability to supply a carbon footprint for the materials; the 300 include third-party manufacturers of finished goods (Unilever, 2024, p. 19). In the pages I read, I did not see how suppliers' views are gathered. Apple acknowledges supplier barriers (Apple Inc., 2025, p. 25) and lists advocacy in Vietnam: in 2017 it joined other companies in urging the government to let businesses buy renewable energy through direct power purchase agreements, and in 2021 it voiced support for a Power Development Plan prioritising clean energy (Apple Inc., 2025, p. 78).

My reading is that a supplier programme passes this condition when the supplier receives something it could not easily obtain alone and the buyer shares the cost. Montiel et al. argue that investment in primary stakeholders such as suppliers benefits the multinational directly and can strengthen positive effects in host-country communities, though they write about subsidiaries, not buyer-supplier programmes (Montiel et al., 2021).

Condition 5: how many goals, and do they pull against each other?

Portfolio discipline means choosing few goals, with reasons, and recording what each choice might make worse. The SDG Compass says not all 17 goals are equally relevant; the first task is to assess impacts along the value chain (GRI et al., 2015, p. 11). Montiel et al. say that 17 goals, 169 targets and 232 indicators can overwhelm firms, and that a multinational need not address every goal: some may focus on the few that fit their mission, culture or strategy (Montiel et al., 2021).

Synergies and trade-offs between SDGs are where a portfolio breaks. Van Zanten and Van Tulder's example, a hypothetical oil company (section 5.3): reforestation to cut emissions can clash with farmland and so limit SDG 2 on hunger, whereas directing it to water-scarce, degraded land adds benefits for water and land (Van Zanten and Van Tulder, 2021). My rule: every SDG taken on needs one indicator, one named owner and one line headed "trade-off".

How do the two documents fare, in the pages I read?

Which claim needs which evidence?

The claims below are hypothetical, written by me for illustration.

This suits a factory willing to answer with real figures and write "not yet" where it has none. It does not suit a team that needs a polished reply in two weeks: condition 3 stops you first. For a small factory, my view is that two goals with a full file beat eight on a slide. If the buyer's concern is wording printed on a pack, read the article on sustainable marketing.

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

The library's origin field names Van Zanten and Van Tulder (2021), Carroll (1991), Freeman and McVea (2001), Banerjee (2008) and Montiel et al. (2021). Van Zanten and Van Tulder: the full open-access text, 18 pages, via an Internet Archive copy. The printed volume runs 3703 to 3720, but I could not open a copy with those page numbers, so I cite sections, not pages. Montiel et al.: the full text on PubMed Central, which has no page numbers, so I paraphrase. Banerjee: the abstract, introduction and passages on the business case in the author's accepted manuscript; I did not read all of it. Carroll: I could not open it and used only the library record and Van Zanten and Van Tulder's account. Freeman and McVea: not read.

SDG Compass: pages 10 to 12, 16, 20 and 22. Unilever's plan: pages 4, 5, 8, 10, 13, 19, 26, 38 and 51 of an Internet Archive copy, since Unilever's server refused me. Apple's report: pages 3 to 5, 9 to 12, 25, 26, 78 and 104 to 107 of 126.