Polle de Maagt

Trend planning.

By Polle de Maagt · First written · 5 min read

Trend planning does not predict the future. It makes sure that when the future arrives, you are not seeing it for the first time.

What it does

Takes loose signals through five steps to a short list of trends, each with a priority and a horizon.

Use it for
Annual planningStrategy workDeciding which trends get budget now
You need
Signals from varied sourcesA few people who know the fieldA 2x2 on a wall
Origin
My guest lecture at Sogang University; horizons adapted from Amy Webb

What is the problem?

A list of trends is not a plan. Without a priority and a horizon, every trend looks equally urgent, so none of them changes a decision.

How does the model work?

Five steps, always in this order:

  1. Research. Collect weak signals. Look for at least three unrelated sources per trend.
  2. Cluster. Group the signals into a handful of bigger trends. One signal rarely matters. The trend ten of them form often does.
  3. Validate. Test each trend with people who work in the field. They kill the trends that only exist in reports.
  4. Prioritise. Place every trend on a 2x2: importance against certainty.
  5. Plan. Give every trend you keep a horizon and a date to look at it again.

The original cycle runs twice a year. I run it every quarter.

ResearchthreeunrelatedsourcesClusterinto a fewbig trendsValidatewith peoplein the fieldPrioritiseimportanceand certaintyPlana horizon foreachagain, twice a year
Researchthree unrelated sourcesClusterinto a few big trendsValidatewith people in the fieldPrioritiseimportance and certaintyPlana horizon for eachagain, twice a year

What does it look like?

The 2x2 asks two questions per trend. Importance: how much does it matter to us? Certainty: how sure are we that it is real?

  • Act now. High importance, high certainty. Plan and budget for it this year.
  • Keep a vigilant watch. High importance, low certainty. Watch for confirmation and prepare options.
  • Informs strategy. Low importance, high certainty. Real, but not core yet. Let it shape the strategy without acting now.
  • OK to dismiss. Low importance, low certainty. Park it until the next cycle.
InformsstrategyActnowOK todismissKeep avigilant watchIMPORTANCE: LOW → HIGHCERTAINTY: LOW → HIGH
InformsstrategyActnowOK todismissKeep avigilant watchIMPORTANCE: LOW → HIGHCERTAINTY: LOW → HIGH
Tactics1-2 yearsStrategy2-5 yearsVision5-10 yearsSystems-level evolution10+ yearsmore certainty ... less
Tactics1-2 yearsStrategy2-5 yearsVision5-10 yearsSystems10+ yearsmore certainty ... less

Then the horizons: tactics for 1 to 2 years, strategy for 2 to 5 years, vision for 5 to 10 years, systems-level evolution beyond 10. The further out, the less certain, so the response gets broader: from actions, to a direction, to watching the system shift.

What did the first run teach me?

My first full run gave standard output: the trends every industry newsletter already carried, stated with more confidence than evidence. The process was fine. The inputs were not. A debrief led to three changes:

  • Look across categories. A trend that shows up in technology, society and creative culture at once is hard to dismiss.
  • Cap the AI trends. If half the board is AI, it is one megatrend in a lot of costumes.
  • Look at the obscure on purpose. Standard sources give standard trends.

Reflection · 3 min readYou still have to know things.AI gives us access to answers in fields we know little about.And one rule: AI collects and structures the evidence, it does not pick the trends. A confident AI trend with round numbers is a yellow flag, not a finding.

How does it run in practice?

This model runs behind Movements. Daily, an agent logs datapoints from my newsletters in an evidence database. Weekly, I review the queue: promote, edit or dismiss. Quarterly, I run the five steps and publish a short read on four or five movements. Each quarter opens with a question: what got stronger, what changed direction, what no longer holds?

The noticing that feeds this uses a more intuitive method: Prime, converge, rationalise.

Why does it still matter?

AI made the scanning manageable. It did not make AI the judge. The work that is left is judgment: what matters, how sure you are, and when to look again.

Try this

Take the three trends your team keeps arguing about. Place each one on the 2x2, give each a horizon, and put the date you will look at it again in the calendar.

Origin: five steps and the 2x2 from my guest lecture on trends at Sogang University. Horizons adapted from Amy Webb, How to Do Strategic Planning Like a Futurist, Harvard Business Review, 30 July 2019. First-run lessons and the Movements workflow: my own practice, 2026.

Your version of this

Tried it, or thinking about it?

Let me know how it went.

Let's talk →How I work
Polle de Maagt

Keep reading

All tools →