Salecare works out how many units, how much revenue, margin and profit the coming months bring. The model is computed in code on your own data, so the same data always gives the same result. Every projected month carries a lower and an upper bound rather than one figure without context.
Sales forecast from the Pro plan. See pricing
Units sold, revenue, margin and costs are summed month by month from the orders synchronised in Salecare, reaching 24 months back. Months with no sales are skipped, so a gap is not counted as a zero. The current unfinished month stays out of the calculation, so it cannot drag the figure down.
The model first works out the typical shape of sales across the year, month by month. After removing seasonality it fits a direction line, that is, it checks whether sales are rising or falling. The shape derived from your own data switches on at 13 months of history, and below that threshold the model uses a typical market profile with a quieter summer and January and a strong November and December.
Every projected month carries a lower and an upper bound, computed from how far past results fell from the trend. The range is drawn to cover roughly four months in five, not every possible outcome. The further into the future, the wider it gets, because the less is known. The upper bound is the safe buffer for a wholesaler order.
The forecast is split into categories by their share of sales in the same period a year ago, and when that period holds no data, over the last 6 months. The categories add up exactly to the forecast total. Products reach the What may run out list by their own sales over the last 90 days, with the combined stock of every warehouse, sales in 30 days, how long the stock lasts and how many units are short. The Download the reorder list button turns them into a file in Downloads.
Order too much and you freeze cash in the warehouse, order too little and the offer leaves the search results in the middle of the season. The forecast in units says how much you will sell in the coming months. The first line gives the projected revenue over the whole period and the month that brings the most of it, the point your stock should be full for. For goods that are easy to sell on you order closer to the upper bound of the range, for slow ones closer to the lower. The What may run out list sets the products' sales against their combined stock from every warehouse, and the Download the reorder list button turns it into a file, so the wholesaler order is assembled on one page. When the forecast turns down, that is a signal not to tie up cash in goods that will sit still over the coming months.
The forecast does not stop at units. The same page shows revenue, margin, profit and costs over the same horizon, so you know how much cash comes in and how much of it stays. A horizon of 1, 3 or 6 months is the one for a conversation about budgets and purchasing limits. The first line gives the revenue of the whole period, and the month table shows each month on its own.
With a seasonal assortment an average of the last few weeks misleads, because November and December follow rules of their own. The model measures seasonality apart from the trend, which is why a Q4 peak month can come out high even on a downward trend. In the month table every month is tagged with its drivers, the Q4 season, holidays and the trend direction. The holiday tag appears when the month covers one of the days that move shopping the most, New Year, 1 or 3 May, All Saints, Easter with Easter Monday or Christmas, together with a window of three days around such a day. The tag only describes the month and does not raise the forecast. Holidays have no factor of their own, so they reach the forecast only through the shape of the year. With less than 13 months of history the shape comes from a typical market profile.
The Allegro panel shows past sales and leaves the conclusions to you. Salecare takes that same history and projects it onto the coming months, with a lower and an upper bound on each of them. Every projected month carries the drivers that shape it, so you know whether a high figure comes from the trend or from the season. The chart sets realised sales as a solid line against the forecast as a dashed line, so one glance is enough to see where this is heading.
An average of the last few months in a spreadsheet mixes the trend with seasonality and says nothing about how far the outcome can differ. Salecare separates the two and states the forecast spread on its own, computed from how far past results fell from the trend. The numbers are produced in code, with no language model involved, so the same data always gives the same result. A spreadsheet has to be refreshed by hand, while the forecast is recomputed every time the page is opened, on the current sales history.
A forecast of turnover alone does not say how much of it stays after commission, shipping, advertising and returns. Salecare projects revenue, margin, profit and costs side by side, computing them from the costs you have recorded in the app. Profit is taken as margin less packing and order handling, account fees and fixed costs, so the figures agree with the rest of the reports and do not have to be assembled in a spreadsheet. With a negative margin the forecast shows a loss rather than a zero.
A forecast over 1, 3 or 6 months, in five metrics, from units to costs.
The forecast is computed from the orders synchronised in Salecare, so it needs a connected Allegro account and a downloaded sales history. Until the first full month of sales the page does not guess and says the forecast appears after it. For the margin, profit and cost forecasts to match reality, purchase prices and costs have to be filled in first. The sales forecast is available from the Pro plan.
A forecast appears even from one or two months with sales, but then there is nothing to measure the spread from, so the model assumes it to be a quarter of the last month's figure and the range comes out deliberately wide. The line under the forecast says how many months of sales it stands on and its spread. With short history the page usually adds the note "Little history, the amounts are rough".
The forecast assumes ad spend stays where it has been, and it knows nothing of planned promotions, price changes or a competitor's moves. A product that has not sold in the last 90 days does not reach the What may run out list. The projected level is capped from above, at two and a half times the highest level in the history once seasonality is taken out, so a short, steep history cannot run away in the projection, which means a sudden jump in sales reaches the forecast with a delay. Salecare orders nothing and changes no offers in this feature, the forecast is a number for a decision, not the decision.
Features that work together with this one.