Two weeks, two uncomfortable findings: a workforce dropped from its own forecast (The Signal), and a consensus you can't take at face value (The Noise).

This week, the answer.

If you can't trust the trend list, what do you trust? Your context. A people strategy that mirrors the top five trends isn't a strategy — it's a summary with a logo. The work is to treat every trend as a hypothesis, test it against your own evidence, and decide — deliberately — what to fund and what to ignore.

That's The Filter — Part 3, and the close of the trilogy. The method that turns ten reports into one decision: yours.

→ Read The Filter

— Tanguy

Link to heatmap → issue #033 (or the PDF download).

The Filter: How to Cut Through the Noise

Part 3 of 3 — a method for turning trend reports into strategy, and a way forward for the profession.

By Tanguy Dulac, Founder & Managing Director, PeopleCentriX

Part 1 called HR value and operating model the filter — the one item on the agenda whose job is to judge the rest. Part 2 showed why it needs to be judged: a consensus that reads more like a market than a measurement. This part is the filter itself.

Porter’s line is usually quoted in the context of products. It applies with more force to the trends on your desk: if strategy is choosing what not to do, a people strategy is defined as much by what it rejects as by what it adopts. Almost no one rejects anything. That is the gap this method closes — three disciplines, preceded by one piece of proof. They are not freestanding tips; they are the People Centricity Methodology™ — Empathise, Strategise, Realise, Analyse — turned on the reports themselves: gather the evidence first, treat every trend as a hypothesis, then define success up front and keep score. Empathise here is not just employee sentiment; it means understanding the context and balancing the needs of every stakeholder, from front-line employees to the board.

Underneath the method is a claim about cause. A trend adopted without evidence intervenes blindly in how an organisation actually works; a trend tested against your own people is one you understand before you act on it. That is why evidence-first is not fussiness but necessity: the filter exists to stop the consensus from disturbing what it promises to improve.

The proof: one audit, objective by objective

The strategy from Part 2 — board-approved, built on 35 interviews and 25 focus groups with executives, managers, and employees — gave us 10 objectives to audit against the publishers’ 2026 consensus, each coded independently by two coders. Mapped to the same 20-theme framework:

Adopt the consensus wholesale and the organisation would have funded priorities its evidence did not support, and missed the ones its mission depends on. The map was half right; nothing in it said which half. The consensus we audited against is published in full as our 2026 HR Trends & Priorities heatmap.

One — treat trends as hypotheses, not priorities

The protection is grammatical before it is analytical. A trend enters your process only as a question: not “skills-based workforce planning”, but does our workforce planning fail because we plan in jobs, not skills? Much of the 2026 corpus reads as instruction — reinvent, transform, unleash — and an instruction can only be obeyed or ignored. A hypothesis can be tested. Reframed that way, a 75%-AI consensus stops being orders to follow and becomes claims to examine against your own reality.

Two — evidence first, always

Hypotheses then face the organisation, not the literature. Interviews, focus groups, workforce data, capability assessments — across employees, managers and executives — are where the consensus is allowed to fail. The rule is non-negotiable: nothing becomes strategy until the diagnosis is validated with the people who will live with it.

It is not a house style but evidence-based practice, carried from medicine into HR: draw on multiple sources, follow an explicit process, weight evidence by trustworthiness (Rousseau and Barends, 2011; Briner, 2024). And here Porter’s discipline bites — welcome the kill rate. If every consensus trend survives your diagnostic, it was a confirmation exercise, not a test. A people strategy that matches the publishers’ top five is not a strategy; it is a summary with a logo.

One caution, because it is the same conflict Part 2 found in the publishers: the filter works only if whoever runs it has no stake in the result. Would this adviser earn anything from the trend being confirmed? If yes, the noise has not been filtered. It has been laundered.

Three — define success before you need it, then keep score

The trends industry has no error-correction loop. In our three-year corpus, not one report revisits its previous predictions. Themes are abandoned, never falsified — wellbeing held a top-ten rank in both 2024 and 2025, then vanished from the top twenty in 2026, with no organisation’s wellbeing problem solved. Every honest forecasting discipline — weather, finance, polling — prices last year’s accuracy into this year’s credibility. HR’s forecasters do not, because their readers have never asked.

Hold your own strategy to the opposite standard. Baseline what it is built to move before you start; define success up front; close each cycle with a review that your board can interrogate. Then score your sources: once a year, ask which of last year’s published trends actually materialised — publisher by publisher. The first review improves your strategy; the second calibrates how much each report has earned for next January.

A method, not a verdict

This series argues that trends should be tested, not adopted. So end where the method begins — with three questions to put to your own agenda: which trends are you obeying as orders rather than testing as hypotheses; where has your own evidence been allowed to overrule the consensus; and how will you know, a year from now, whether you were right.

Within three years, the capability that distinguishes leading HR functions will not be how fast they adopt the consensus, but how much of it they can show they were right to reject.

A larger shift sits underneath all this. A People function that imports the consensus is a curator: it arranges what others have already decided. A People function that produces its own evidence is something else — a discipline, with the Chief People Officer as its clinician, who diagnoses before prescribing and checks whether the treatment worked. That is the distance between being briefed on the future and being equipped to author it; and between a cost centre and a voice the board cannot do without.

This series opened with a person — the one in nine the forecast wrote off, the wellbeing line it let go dark, the colleague whose pay, health and prospects are the whole point of this profession. The filter is how you put the mirror down, look at the room, and write that person back in: every trend you test against your own evidence asks not is this what everyone is doing? but is this what our people actually need?

One question for your next planning cycle: when did your organisation last reject a trend — on evidence?

Mini-lexicon

The filter — HR value and operating model used as a discipline: the judgement that decides which trends fit one organisation.

People Centricity Methodology™ — PeopleCentriX’s evidence loop — Empathise (understand the context and weigh every stakeholder’s needs, from front line to board), Strategise, Realise, Analyse.

Trend as hypothesis — treating each trend as a question to test against your own evidence, not an instruction to obey.

Evidence-based practice — drawing on multiple sources, following an explicit process, and weighting evidence by trustworthiness (Rousseau and Barends; Briner).

The kill rate — the share of consensus trends a diagnostic rejects; a healthy filter welcomes it.

Source audit — scoring each publisher’s predictions once a year against what actually materialised, so credibility is earned, not assumed.

Method

The filtration method described here is the People Centricity Methodology™ — Empathise, Strategise, Realise, Analyse — a continuous evidence loop used in PeopleCentriX engagements.

References

Briner, R. (2024). Evidence-Based HR: A New Paradigm. Corporate Research Forum.

Rousseau, D.M. and Barends, E.G.R. (2011). “Becoming an evidence-based HR practitioner.” Human Resource Management Journal, 21(3), 221–235.