Market PredictionsBeta
Zalando goes on sale in week 40. You know it six weeks earlier.
Market Predictions in Reyo forecasts which competitor starts which campaign in which calendar week: Reyo learns every competitor’s rhythm from the newsletters, push messages and ads of past years and turns it into a probability per calendar week. 4 to 24 weeks ahead, 84% hit rate in backtesting (four-week horizon, beta).
The market is not random. It repeats itself.
Retailers send in patterns. Sales start on the same weekdays, mid-season promotions return in the same weeks, and Black Week announcements arrive earlier every year by roughly the same margin. Anyone who has read 50,000+ newsletters and 35,000+ Meta ads from 300+ brands since December 2023 can see these patterns. Reyo has – and it has trained a model on them.
For every competitor you watch, Reyo estimates the probability that a campaign goes live in a given calendar week, and where the signal is strong enough, what kind: sale, promotion, special or content. The result is a matrix of competitor × week that you read like a calendar – with the confident cells filled in and the uncertain ones left visibly empty.
Monitoring tools show you what happened last week. Predictions tell you what is likely to happen in week 43, so that you plan your own campaign around it instead of reacting to it on Monday morning. That is the difference between reading the market and being early.
Three views on the next eight weeks.
Every prediction comes with its own confidence, so you always know how much weight to put on it.
Competitor × calendar week matrix
One row per competitor, one column per week. Each cell carries a probability and, when the model is confident, the expected campaign type. Cells below 30% stay empty rather than pretending to know.
Quality traffic light
Activity, type and timing are rated separately. Activity is reliable at 84% hit rate in backtesting; the campaign type is more cautious and says so. You see the rating next to the matrix, not in a footnote.
Quietest week
Reyo marks the week with the least expected competitor activity and says what that means for you: room for your own campaign, with fewer sales competing for the same inbox.
Honest about the horizon: the first four weeks are where the model is strongest. From week five onward the forecast leans mostly on each competitor’s base rate – useful for rough planning, not for a launch date.
From forecast to decision
A probability is a number. A recommendation is a decision.
Predictions do not stop at the matrix. When the model expects a discount wave in week 40 and your own autumn sale is planned for the same week, Reyo turns that into a recommendation: move it forward, adjust the discount, or hold – with the forecast, the past campaigns and the external signals attached as evidence.
That recommendation lands in your Monday digest and, with one click, in the Planner. The forecast becomes part of your calendar, not a separate report you have to remember to check.
What Reyo says about it
A forecast is only useful if it changes a plan.
In week 39 the inbox is almost yours: 47% expected activity on average, the lowest value in the eight-week horizon. Zalando, About You and H&M are expected to start their autumn sale only in week 40 (94 / 83 / 58%) – whoever sends one week earlier meets far fewer discounts. If your autumn campaign can move by seven days, this is the week.
- 01Activity forecast week 39: 3 of 8 watched competitors below 40% (Mango 37%, Tom Tailor 29%, Hunkemöller 22%), lowest value of the eight-week horizon
- 02Base rates: Zalando 86%, About You 72%, Peek & Cloppenburg 52% weekly campaign probability
- 03Past two years: autumn sale wave started in week 40 both times
Sample recommendation from the demo tenant. Names are public brands watched as a market, not customers.
What people ask about Predictions.
See the next eight weeks of your competitors.
In the demo we open the matrix for the brands you name – live, with the current quality rating.