Category-Based Demand Forecasting
A category-based demand forecast is a signal that shows which products in a category are more likely to see increased movement in the near term. Hermes calculates this probability from historical price and ranking data. The output is a not a guarantee: it does not give sales volume; it only prioritizes which products should be monitored more closely.
Where the forecast comes from
The indicator is based on the trend over time of three movements visible from outside: price changes within the category, product position shifts in ranking, and changes in stock visibility. When these three are read together, periods when movement concentrates or calms down in a category can be distinguished.
The logic is simple: if a product's ranking is moving up, several sellers are cutting prices at the same time, and stock visibility is shrinking quickly, a probability of rising demand for that product forms. Hermes calculates this probability; it shows the direction and relative intensity of demand, not its size.
Reading the probability indicator correctly
The indicator is not a note; it is a prioritization tool. A high probability does not mean "buy this and sell it"; it means "watch this product's movement closely." A low probability does not mean the product is bad; it means there is no clear movement in the data that can be read from outside.
The value of the signal emerges not from a single measurement but over a time series. If your scan frequency is hourly, you see intraday fluctuations; if daily, you see more the direction of the trend. Making stock decisions based on a high or low signal before watching the same product for a few weeks is premature.
Combining the signal with your own data
A demand signal alone does not produce a commercial decision; you need to add the cost side. You can see net profit per product after commission, shipping and advertising ProfitPulse , and for a quick preliminary calculation use profit margin calculator tool.
The second layer is seasonality and the campaign calendar. Part of the movement Hermes sees comes not from real demand but from marketplace campaigns. If you know how the category behaved in past periods, it becomes easier to separate what part of the signal is campaign effect.
How to do it step by step?
1. Narrow the category
Work with the sub-segment you actually sell in, not the parent category. In a broad category the signal averages out and loses its discriminating power.
2. Set a reference period
Use the last few weeks as the baseline. Without a comparison period there is no ground against which the term "high probability" can be compared.
3. List the high-probability products
Put the products in the upper band of the indicator on a separate watchlist. This list is not a purchase list; it is a review list.
4. Check price and ranking history
For each product, check when the movement started. Price tracking history lets you tell whether the signal was created by a campaign or by a lasting trend.
5. Cross-check with profitability
Calculate whether the product that appears to be in demand leaves you net profit under your cost structure. If it does not, a high signal changes nothing.
6. Set up alerts
Turn on instant price change alerts for the products you choose. That way, instead of opening the indicator by hand every day, you are notified when movement happens.
Signal combinations and possible interpretations
| Visible movement | Possible interpretation | Verification step |
|---|---|---|
| Ranking rising, price stable | Organic interest may be increasing | Rankings of other sellers of the same product |
| Several sellers cutting prices at once | A campaign period may have started in the category | Marketplace campaign calendar |
| Stock visibility decreasing, price rising | Supply may be tightening | Price and availability on the supply side |
| Ranking falling, price falling too | Interest may be weakening | Is the same trend present across the category? |
| New sellers entering the category | Competition increasing | Lower bound of the price range |
| Price and ranking unchanged | No readable signal from outside | Put the forecast on hold and wait for data to accumulate |
Common mistakes
1. Using the demand forecast as an order quantity. The indicator is a probability, not a quantity. Making order decisions based on the signal means handing over the responsibility for remaining stock entirely to a forecast.
2. Making decisions by looking at one day's signal. A one-day movement is often a campaign, a weekend effect or a single seller's test. A trend becomes reliable only over a series of several weeks.
3. Following the signal without a profitability check. A product whose demand is rising may still not make you money after commission and shipping. High-volume losses grow faster than low-volume losses.
4. Mistaking the signal for a competitor's sales volume. The indicator is derived from publicly visible price and ranking movements. A competitor's actual sales volume is not in it, and it is not available in any external tool either.
Limitations: what does it not cover?
Applying these with awareness lets you set realistic expectations:
A demand forecast does not give sales volume. The output is a probability band. It does not produce a number of the kind "this much will be sold of this product"; a tool that tries to produce that number is estimating it.
A competitor's internal data cannot be seen. Sales volume, revenue, profit margin and supply cost cannot be measured with any external tool. This data is only in the seller's own panel.
External events are not part of the model. Regulatory changes, currency moves, supply chain disruptions or a social media wave do not show up in historical price and ranking data; the signal reacts with a delay after these events.
Scan frequency sets the ceiling on accuracy. In a daily scan, campaigns that open and close within the day are not recorded. Hourly scanning narrows this gap but does not fully close it.
Signal is weak for new products. There is no series to calculate for a product without historical data. For newly listed products, the indicator should be read with caution until enough data accumulates.
Frequently asked questions
How far ahead does the demand forecast look?
It measures how past movement carries into the recent period. In practice it is meaningful over a window of a few weeks; as the horizon extends, the explanatory power of the signal drops quickly.
How often is the indicator updated?
Hermes's scan frequency is hourly or daily depending on the subscription. The indicator refreshes on this rhythm; intraday fluctuations are not captured in a daily plan.
Does Hermes change my price automatically based on the forecast?
No. Hermes measures and alerts. You always make the price change yourself; there is no automatic price intervention.
Can I see how many units a competitor sold?
No. What can be read from outside is price, ranking and stock visibility. Internal data such as sales volume and profit margin cannot be seen with any external tool.
Which marketplaces does it work on?
Trendyol, Hepsiburada and Amazon Turkey. For category-based reading, product analysis approach strengthens the context of the signal.
Doing it with the tool
A category-based demand forecast is a layer that speeds up decisions rather than making them. The right way to use it: build a short watchlist from the indicator, verify the list with your demand forecasting approach and your own sales history, then ProfitPulse look at the net profit side.
The indicator's Hermes If you want to try how it looks within or request a you can use it, and see how scan frequency changes with the plan on the page or page.
Last updated: September 18, 2026