How can AI support range and line planning in fashion?
Range planning decides how many styles, at which prices and in what depth a collection should have. AI can forecast new items and test scenarios, but the creative and strategic call stays human.
KEY TAKEAWAYS Summary by the editors
- Range and line planning translate a financial plan into a structured collection: number of options per category, price architecture, newness versus carry-over, and depth per option.
- AI contributes mainly through forecasting demand for new styles from attributes, images and similar past products, and through fast scenario comparisons of different range structures.
- Research on the VISUELLE dataset from Italian fast-fashion company Nunalie showed that adding Google Trends signals to image and attribute data improved new-product sales forecasts.
- Mango (2025) and Marks & Spencer (2024) have both announced integrated planning programmes covering assortment or range planning alongside financial and demand planning.
- Models learn from past ranges, so they tend to favour what has worked before; deliberate space for new directions has to be set by people.
AI can support range and line planning by estimating demand for each proposed style, including new ones without sales history, and by letting planners compare alternative range structures quickly against financial targets. It works best as an analytical partner in the range review: it flags overlaps, gaps and over-deep options, while designers and merchants decide the creative direction.
What is range planning in fashion?
Range planning (often called line planning at brands and assortment planning at retailers) is where a financial plan becomes a collection. Planners decide how many options each category should have, how they spread across price points, how much is new versus carry-over, which colours are offered, and how deep each option is bought. The output is a range plan or line sheet that guides design, sourcing and buying.
The difficulty is that most of these choices are made months before the season, for products that often do not yet exist. Traditional planning relies on last season's results, category reviews and experience, typically in large spreadsheets that are hard to update when the plan changes.
Because range decisions fix the frame for every later step, errors here are expensive: an option that should never have been in the range consumes design time, sampling cost, minimum order quantities and store space, and usually ends in markdowns.
Where in range planning can AI help?
| Planning step | AI contribution | Human role |
|---|---|---|
| Category and option count | Projects sales by category from history and trends; tests scenarios | Sets strategy, brand positioning and targets |
| Price architecture | Analyses sell-through and price sensitivity by price band | Decides price positioning versus competitors |
| New style forecasting | Estimates demand from attributes, images and similar past items | Judges design quality and trend relevance |
| Range overlap and gaps | Finds similar options competing for the same demand; flags white spaces | Decides which to cut or develop |
| Depth and buy quantities | Converts forecasts into buy depth per option and size | Approves risk level per option |
| Localisation | Suggests which options suit which store clusters or markets | Balances local fit against complexity |
How does AI forecast new styles without history?
The common approach is to describe each new style by attributes (category, silhouette, colour, fabric, price) and, increasingly, by its image, then find past products with similar characteristics and learn how they sold. External signals can be added. A study published in the Journal of Forecasting used the VISUELLE dataset, which contains 5,577 new products sold between 2016 and 2019 by the Italian fast-fashion company Nunalie, with images, metadata and sales. The authors combined product images and attributes with Google Trends data and found that the external trend signal improved forecast accuracy compared with leaving it out.
The gains in such research are real but modest, and forecast error for new items stays far higher than for carry-over basics. That is why AI range tools are most useful for ranking options and sizing depth (which options deserve a deeper buy, which should be tested shallow) rather than for producing a single reliable number per style.
Which companies are investing in AI range planning?
Published examples tend to be large retailers replacing older planning tools. In March 2025 Mango announced it would modernise merchandise financial planning, assortment planning and demand planning with an integrated, AI-powered planning platform, aiming to replace fragmented, outdated systems. Mango had already described a wider AI programme: Glossy reported in 2024 that the company ran 15 AI platforms, including tools for pricing and a generative AI platform for design and product teams.
In January 2024 Marks & Spencer announced a three-year programme for its clothing and home business, starting with merchandise planning, sales, stock and intake, and range planning, followed by forecasting and replenishment. The company said an end-to-end platform would let it better use data and AI to inform decisions.
What data is needed?
- Structured product attributes for past and planned styles, filled consistently (a fuzzy attribute such as "casual" is far less useful than silhouette, fabric and fit).
- Product images linked to the article master, if image-based similarity is used.
- Sales, stock and markdown history by option, so the model learns full-price performance, not just units.
- The financial plan by category and period, to keep the range consistent with budgets.
- External signals where available, such as search trends, used with caution and tested for real predictive value.
What are the risks of AI in range planning?
- Homogenisation: models reward styles similar to past winners. Without deliberate space for newness, ranges can drift towards the safe middle.
- False precision: a forecast to the unit for a new style suggests certainty that does not exist; ranges and confidence levels are more honest.
- Data leakage from bias: options that were under-bought or poorly displayed look weak in history and get cut again.
- Process fit: range reviews involve design, merchandising and sales; a tool used by only one team rarely changes decisions.
How should a team introduce AI into range planning?
Begin with analysis of the last two or three seasons: which options were redundant, which price bands underperformed, where depth was misjudged. Use that to build trust in the data. Then use the model in one category's range review as a second opinion, documenting where planners disagree and checking at season end who was closer. Over time the model can take on more of the arithmetic, while the range review focuses on brand direction and risk appetite.
The aim is a range plan that is faster to build, easier to change and explicit about risk, not one that is written by an algorithm.
Frequently asked questions
What is the difference between range planning and assortment planning?
The terms overlap. Range or line planning usually describes building the collection structure at a brand, while assortment planning often describes deciding which products go to which stores or channels at a retailer. Both decide breadth, price architecture and depth.
Can AI predict which new fashion styles will sell?
It can estimate demand for new styles from attributes, images and similar past items, and research shows external signals like search trends can improve accuracy. Error remains high, so forecasts are best used to rank options and set depth.
Does AI make fashion ranges more similar?
It can, because models learn from past winners. Teams counter this by reserving part of the range and budget for tests and new directions that are judged by people.
What is a line plan in fashion?
A line plan is the structured outline of a collection: number of styles per category, price points, colours, delivery drops and planned quantities. It links the financial plan to design and sourcing.
One edition every weekday morning. Read in five minutes. Free for industry professionals.
SOURCES
- arXiv: Well Googled is Half Done: Multimodal Forecasting of New Fashion Product Sales with Image-based Google Trends
- Retail TouchPoints: Mango Ramps Up Data-Driven Planning for Global Operations
- Glossy: How AI is fueling Mango's $3.4 billion business
- Just Style: M&S to digitally transform clothing and home end-to-end planning system
