Every autumn I hear the same story from retailers in different sectors: the buying team ordered by gut, the shop sold out in week two, the website sat with stock nobody wanted, and the marketplace listings went dark at the exact moment demand peaked. Seasonal planning in a multichannel business is not one forecast, it is a set of decisions about who gets which stock and when. The retailers that do it well treat it as a process with a calendar, not a spreadsheet with a hope.
Why seasonal planning is harder with many channels
A single shop retailer plans a season with one demand line and one set of markdown levers. Add a website and marketplaces and the same unit of stock has three customers bidding for it: the shopper who wants it today, the online customer who wants it this week and the marketplace buyer who wants it with a Prime badge. Each channel has different pricing, different margin and different consequences for running out. The planning problem stops being "how much" and becomes "how much, where, and when can we move it".
The Internet Retailing coverage of UK multichannel operations makes the same point from the demand side: the channel mix shifts seasonally and nobody can predict it precisely, so the plan needs slack in the right places, which is a stock control question, not a demand question.
The demand baseline comes first
The foundation is a baseline: what did this season sell last year, by week, by channel, by SKU, net of returns and markdowns. Most retailers have this data scattered and dirty. The accuracy discipline article explains why the record is unreliable in the first place, and the seasonal plan inherits every error. Fix the record before you plan, or you are planning against last year's fiction.
Adjust the baseline for three factors: like for like growth, known changes in the range, and channel shifts you can already see. Then write the plan in weeks, not in total units, because the season is a curve and the curve is where the stock gets trapped at the end.
The channel split is a decision, not a forecast
The single most useful conversation in seasonal planning is deciding which channel gets priority on scarce stock. It is a margin and strategy decision, and it should be made by the commercial team, not discovered by the system. A fashion retailer may protect the flagship store because full price sell through there is strongest. A homeware retailer may protect the website because that is where the marketing spend lands. A marketplace first business protects the listings that drive its seller ranking.
Once the priority is decided, the ERP enforces it with allocation rules and per channel safety stock, the mechanics covered in preventing overselling. The rule matters less than the fact that someone decided it. Retailers who skip the decision get the default, which is that the first channel to sell wins, and that is rarely the channel you wanted to win.
Lead times and the last order date
Seasonal stock arrives on a schedule, and the schedule has a last order date beyond which nothing new will arrive in time. Work backwards from the peak week: manufacturing or supplier lead time, sea or road freight, customs, inbound processing, quality checks, putaway. Every week counts. The Retail Times trade reporting regularly covers supply chain disruption stories from UK retail, and the pattern is always the same: the stock that misses the last order date does not miss the season, it arrives for the markdown.
Plan two waves where you can: an opening order sized to confirmed demand and a top up order sized to the sell through in the first weeks. Two waves beat one big bet, because the second wave uses real data from the season you are actually running.
The weekly review that saves the season
The best seasonal plans are wrong by week two, which is fine if the review cadence exists to correct them. The weekly seasonal review answers five questions: what sold ahead of plan, what sold behind, what is the sell through rate per channel, where is stock trapped, and what moves do we make now. The moves are the point: reallocate from the slow channel to the fast one, pull forward the top up order, mark down the trapped item before it is worth less than the label says.
In the ERP this means the stock movement and allocation tools, not more spreadsheets. The one business model describes the operating model where these reallocations happen in hours instead of weeks.
What the ERP should be doing
By the time the plan is written, the ERP's job is execution: hold the allocation rules, publish availability to each channel from one record, flag slow movers, and report sell through by week and channel without manual assembly. The marketplace execution article shows the channel side of that execution, and the one stock record article shows the margin side. A seasonal plan executed on fragmented systems becomes a series of reconciliations during the busiest weeks of the year, which is exactly when nobody has time for them.
The closing thought from client work is about expectations. A good seasonal plan does not eliminate markdowns or stockouts, it concentrates them where they are cheapest and most visible. The stockout cost framework helps you decide which of the two you would rather have, and the answer is usually "a controlled markdown on a visible line", which is a decision you can only make when the stock data is honest enough to trust.