Guide

How to build a trend board

A working method for building a trend board that survives contact with a buyer — how to gather signals, cut them down, name the story, and defend it in the room.

By YakLab · Published 26 August 2026 · 8 min read

A trend board is not a mood board

The two get used interchangeably and they are not the same object. A mood board captures a feeling. It has to be seductive and nothing else. You can build a beautiful one from images you like, and nobody will ask you to justify it.

A trend board makes a claim about the future: that something is moving in a particular direction, for a particular customer, in a particular season. That claim will be used to commit money — to a fabric order, a colour card, a range plan. So it has to be persuasive and defensible, and the second half is where most boards fall down.

The practical difference is evidence. A mood board carries images. A trend board carries images plus a reason each one is there.

Start with the decision, not the images

Before you open a single tab, write down the decision the board is meant to inform. Not the topic — the decision. “Which three colour stories go into the SS27 womenswear range” is a decision. “Spring colour” is a topic, and a board built from a topic will drift until it means nothing.

Then fix four things in writing:

• Customer. Not a demographic bracket — a person with a wardrobe and a budget. “Women 25–40” tells you nothing. “Buys three or four considered pieces a season, shops mid-market, dresses for an office that stopped enforcing a dress code” tells you what to reject.
• Category. Outerwear reads a trend differently to jersey. Be specific.
• Season and horizon. Which delivery, and how far out. See lead time.
• Stance. Are you looking for what is emerging, what is already commercial, or what is on the way out and needs killing? These produce completely different boards from the same research.

Written down, that paragraph is your filter. Every image either serves it or goes.

Gather signals from more than one direction

The most common failure in trend research is monoculture: forty images that all came from the same corner of Instagram, arranged to look like consensus. If every signal shares a source, you have found an algorithm's preference, not a trend.

Pull deliberately from different kinds of source, because each lies in a different way:

• Runway and collections. Directional and early, but heavily filtered through a small number of houses and stylists. Tells you where the conversation is, not where the money is.
• Retail assortment. What is actually hanging, at what price, in what depth. Slow to move but honest — depth of buy is a real commitment by someone with a budget.
• Resale and secondhand. An underused signal. What people are searching for and paying over the odds for tells you about demand without a marketing budget behind it.
• Street and social. Fast, noisy, and prone to feedback loops. Useful for velocity, dangerous for direction.
• Materials and colour authorities. Yarn and fabric fairs, mill development, colour cards. These move earliest because they have to.
• Outside fashion entirely. Interiors, product design, film and television art direction, food, music. Colour and surface migrate across categories, usually into fashion rather than out of it.

Record where each image came from and when. It feels like admin. It is the thing that lets you answer “why do you believe that?” six weeks later in a room full of people who are about to spend money.

Cut hard, then cut again

Gathering is the easy half and the half that feels productive. The value is in the cut.

Work in two passes. The first pass is by relevance: does this serve the customer, category, season and stance you wrote down? Anything that does not, delete — not “move to a maybe folder”, delete. The maybe folder is where boards go to become mush.

The second pass is by redundancy. Where five images are making the same point, keep the one that makes it most clearly and lose the rest. Repetition feels like weight of evidence and reads as padding.

Most working boards land somewhere between eight and fifteen images per story. Past that, the eye stops reading individual images and starts reading texture — and the argument gets weaker, not stronger.

Name the story

A trend board without a name is a folder. The name is what people repeat in the corridor afterwards, and what stops the story drifting as it moves through design, buying and marketing.

Good names are short, concrete and slightly odd. They describe a world rather than a product. Avoid the season-agnostic vocabulary that could be attached to anything — “Modern Minimal”, “Urban Edge” — because a name that could belong to any story tells nobody anything.

Underneath the name, write three things:

• A one-line tagline — the story compressed to a phrase.
• A short summary, sixty to ninety words. If you cannot get it into ninety words, the story is still two stories.
• An anti-brief: what this story is not, and which neighbouring idea it keeps being confused with. This single line prevents more drift than anything else on the board.

Build the palette last

Palettes built first become a constraint you spend the rest of the process fighting. Built last, they are a summary of what you actually found.

Pull colours from your surviving images rather than from a swatch library, then discipline them: a working palette is usually six to ten colours with a clear role for each — two or three grounds, a mid-range, one or two accents, and a neutral that lets the rest breathe. Name every colour. Unnamed hex codes get substituted by whoever handles them next.

Sense-check the palette against the category. A palette that works in knitwear may be unachievable in a technical outerwear fabric at the price you need, and a colour that sings on screen can die in a matte cotton.

Present it as an argument

The order that works, in a room:

1. The claim. One sentence. What is moving, for whom, by when.
2. The evidence. Your images, grouped by source type rather than by prettiness, so the range of your inputs is visible.
3. The palette, named.
4. The implication. What this means for the range — shapes, fabrics, price points, which existing bestseller it threatens.
5. The risk. What would have to be true for you to be wrong, and what the early warning sign would be.

That last point is the one people skip, and it is the one that earns you credibility. A forecaster who names their own downside is treated as an analyst. One who never does is treated as a stylist with opinions.

Five ways trend boards fail

• Confirmation shopping. Deciding the story first and collecting images to support it. Build the pile before you name it.
• Single-source monoculture. Everything from one platform, one season of shows, one retailer.
• No horizon. Mixing signals that are eighteen months out with things already in-store, so nobody can tell what to act on.
• Too much of everything. Forty images, twelve colours, three overlapping stories. Reads as indecision.
• No provenance. Nobody can trace where a signal came from, so the board cannot be defended and cannot be updated.

---

None of this is complicated. It is just slower than it looks, and the slow parts — the gathering, the cutting, the writing down of sources — are the parts that make the board worth anything.