↑↓ selectEnter openEsc close

    1. Elbuz

    How to read supplier demand analysis metrics

    11 min readUpdated

    For each supplier product, Elbuz shows eight demand metrics — from units sold to when stock runs out. Each backs a decision: what to buy, what to wait on, what not to trust.

    When you need this:

    • Choosing what to buy — sort by “Rating”: top demand first.
    • High sales, but all at once — low “Stability” means a batch was bought out, not steady demand.
    • Supplier running out — “Days” under 7 turns red with “⚠️”.
    • Stock jumps — an “Anomaly” triangle says check the chart first.

    Where to start: open the demand analysis window — it's already sorted by “Rating” — and review the top rows.

    What the window shows

    • “Sold” and “Sell-through %” — how many units the supplier let go during the period and what share of what they had.
    • “Rating” — a 0–100 score for sorting: the top of the list is your purchase candidates.
    • “Trend” — whether demand is growing, falling or holding steady.
    • “Stability” — tells steady demand apart from one large buyout.
    • “Days” — how many days until the supplier runs out of stock.
    • “Anomaly” — warns that stock jumps around and the row's numbers deserve less trust.
    • Settings — the period, rating weights and advertising thresholds, shared by all sources.

    What units sold and rating mean

    One row is one supplier product. By default, rows are sorted by “Rating”, highest first. Below the filters, Elbuz shows a short hint on the columns; here's what each number actually calculates and what decision follows from it.

    Grid indicator columns: Sold, Sell-through, Rating, Trend, Stability, Days, Advertise
    1Rating bar
    2Trend label
    3Days value
    4Advertise icon
    1. The rating bar fills in proportion to the product's demand score.
    2. The trend label with a colored dot shows whether demand is hot, growing, stable or declining.
    3. The Days column shows how many days the remaining stock will last, with a warning icon when it is short.
    4. The megaphone icon marks products recommended for advertising.

    “Sold”

    The sum of all stock decreases between neighboring history points during the period. The supplier had 100, then 80, then 60 — 40 sold. Restocks don't count: if stock grew between two points, the increase is simply ignored.

    Decision: this is the upper limit of how many units you could have sold if you had this product. It's the only number in the window you can read literally — rely on it when the other metrics seem to contradict each other.

    “Sell-through %”

    Units sold divided by everything the supplier had in stock during the period: starting stock plus all restocks in the period. So it's a true share of what sold out, and it never goes above 100 %.

    The column uses the same color scale as “Rating”: 70 % and up is green, 40 % and up is yellow, lower is red.

    Decision: 70 % means the supplier let go of most of what they had during the period; 100 % means the item sold out completely and there's no stock left at the last point.

    “Rating”

    A combined score from 0 to 100, drawn as a bar: green from 70, yellow from 40, red below that. It has three parts, each with its own weight from the settings:

    • sell-through rate — “Sell-through %”, weight 0.5;
    • stability — weight 0.3;
    • sales velocity — average daily units sold, multiplied by 10 and capped at 100, weight 0.2. Ten units a day already give this part the full 100 points.

    The three weights are normalized by their sum, so the 0–100 scale doesn't depend on the numbers you enter: only the ratio matters. Whole units and tenths give the same result.

    Decision: the rating is made for sorting, and it does that well. The top of the list gives a short list of items to consider before buying, instead of comparing hundreds of rows by eye. Don't compare scores across suppliers: with large stock, the same sales give a lower sell-through percentage and a lower score.

    How to read trend, stability, days and anomaly

    “Trend”

    One of six words, with an icon: “Hot”, “Growing”, “Stable”, “Declining”, “Cold”, “New”.

    Elbuz checks the conditions in order: fewer than three points — “New”; rating below 20 — “Cold”; fewer than four points — “Hot” at a rating of 70+, otherwise “Stable”. Only from four points is the history split in two halves, and the second half's sales compared with the first's.

    Elbuz compares the halves' average daily units sold, within the “Analysis period” only. “Hot” comes first: a rating of 70+ and sales growth over 20 %.

    Then a difference over 15 % one way gives “Growing”, the other way “Declining”, anything in between “Stable”. The trend doesn't depend on the number of points or on how often the price list arrived: both halves are converted to calendar days.

    Conditions are checked top to bottom, and the first one that matches sets the word in the column. The left branch is a three-point history where the halves aren't compared yet; the right one is four points or more
    YesNoYesNoYesYesNoNo, four or moreYesNoup more than 15 %down more than 15 %within 15 % either wayFewer than three history pointswithin the “Analysis period”?“New”“Rating” below 20?“Cold”Exactly three points?“Rating” 70 or higher?“Hot”“Stable”Average daily sales, second halfvs first: “Rating” 70 or higherand growth over 20 %?How much did saleschange in the second half?“Growing”“Declining”
    • Action or result
    • Decision: the path depends on the answer

    Decision: “Cold” and “New” are reasons to leave the row alone for now: in the first case there's too little movement, in the second there's no data yet. “Declining” on an item you buy is a reason to check whether the supplier is clearing out stock.

    “Stability”

    The share of price list imports after which stock went down: three decreases in five intervals between points gives 60. These aren't days. If the price list arrives every three days with lower stock each time, stability is 100, though stock moved on five days out of fifteen.

    Decision: this metric tells steady demand from one large shipment. High “Sold” with low “Stability” means the product went in one go — someone bought a batch, and it may not repeat; don't buy on that signal.

    High “Stability” means the product sells evenly and you can stock it permanently. But infrequent imports inflate the number: compare items within one price list, not price lists with different schedules.

    “Days”

    A forecast of when the supplier's current stock runs out: stock divided by average units sold. The average is per calendar day, not per import, so it doesn't depend on how often the price list arrives.

    Colors: under 7 is red with “⚠️”, under 14 yellow with “⏳”, the rest green with “✓”. A dash “—” means nothing sold at all or stock is already zero.

    Decision: a red “⚠️” means the supplier is about to run out. If you need the product, order now, not next week; a green “✓” means no rush. Read a dash with the “Quantity” column: zero there means the supplier is already out.

    “Advertise”

    A green megaphone “Can advertise” or a gray minus “Not recommended for advertising”. The megaphone follows rating, days of stock and stability thresholds and goes to the product card. Details: which products get the mark.

    “Anomaly”

    The last column: a yellow triangle or a dash. The triangle means the item's stock jumps — between neighboring points it changed by over half the average stock, or went “0 → in stock → 0”. An item with under three history points is never an anomaly.

    Rating and trend for such a row are calculated as usual, but trust them less: the difference between two imports here is more likely a stocktake or technical upload than demand.

    Decision: a triangle means open the product card and check the chart before deciding anything from the row's numbers.

    What else is in the window

    Above the grid is a row of stat tiles: “Products”, “Average rating”, “Sold in period” and the number of items in each of the six trends. Trend tiles are clickable: a click filters the grid by that trend, a second click clears the filter.

    While the filter is on, a chip with the trend name sits on the right of the filter row — its cross also clears the filter.

    With no rating rows, the tiles stay with zeros, and one of two hints appears below the filters: “Select a price list for analysis” (no price list selected yet) or “The rating has not been calculated. Click the “Recalculate” button.” (a price list is selected, but has no rows).

    Supplier Product Demand Analysis window: trend statistics, toolbar, filters and the first grid rows
    1Statistics row with trend counters
    2Recalculate button
    3Analytics button
    1. The statistics row shows the number of products and how many fall into each trend; clicking a tile filters the grid by that trend.
    2. The Recalculate button rebuilds the demand rating from the stock history.
    3. The Analytics button opens the demand analytics view.

    The toolbar has two labeled buttons on the left and four icons on the right:

    • “Recalculate” — recalculates the metrics from the history already collected for the selected price list;
    • “Analytics” — comparison with your own sales;
    • question mark — the “How to read the columns” hint;
    • “Export” — exports the rating to a file;
    • “Refresh” — reloads the grid and the stats;
    • “Settings” — opens the window settings.

    Filters below the toolbar: “Supplier” and “Price list” (only when the window is open in all-suppliers mode), “Category” with the product count in brackets, and “Search” — by SKU, name and manufacturer; it searches shortly after you stop typing, no need to press anything.

    The “Supplier” list shows only suppliers that already have history, while the “Price list” list shows all active sources of the selected supplier, including those without the checkbox selected; such a price list gives an empty grid.

    The grid itself has a filter row — it works on SKU, name and category. You can't filter by the “Trend” column there — use the trend tiles in the stats row instead. There's no separate period filter in the window: the period comes from the settings. Rows load in batches of 50.

    There are eighteen columns. Eight of them are the metrics covered above; the rest:

    • the first, unlabeled column — a plus if the product isn't in your catalog, and a gray check mark with the hint “Product already in catalog” if it is;
    • “Status” — a red NEW badge for items with no match in the catalog, otherwise the product number like “ID: 1234”;
    • “#” — the product's place in this source's rating, “# in cat.” — its place in its category's rating;
    • “SKU” and “Name” — as sent by the supplier; clicking the name opens the product card with a chart;
    • “Category” — the category from the supplier's price list, not yours;
    • “Quantity” and “Price” — from the last history point;
    • “Dynamics” — between them, a sparkline of stock over the period; hover to see what the stock changed from and to. If there are fewer than three history points, the cell is empty.

    Product card

    Clicking the name opens a window with three parts. On the left — SKU, “Manufacturer”, “Category”, “Quantity” and “Price”, taken from the current price list row. On the right — a chart with two lines: quantity on the left axis, price on the right, with the product's whole history as points.

    Below is the “Metrics by period” table with units sold, percentage, stability, rating and trend; it almost always has one row, because one calculation is stored per product. If the product has no history, you'll see “No data for analysis” instead of all this.

    Product card with stock and price history chart and the metrics by period table
    1Quantity line
    2Price line
    3Metrics by period table
    1. The quantity line shows how the supplier's stock decreased over the period.
    2. The price line shows supplier price changes over the same period.
    3. The Metrics by period table lists sold units, sell-through, stability, rating and trend.

    Window settings

    The “Settings” icon on the toolbar opens twelve fields. The key point: the settings are shared by the whole window, not per source. Change a weight or threshold, and it applies to all suppliers at the next calculation.

    Settings dialog of the Supplier Product Demand Analysis window
    1Analysis period (days)
    2Min quantity for analysis
    3Three weight fields
    4Update “Can advertise” flag in product catalog
    1. Analysis period (days) — how many days of stock history the rating is calculated over.
    2. Min quantity for analysis — products with a lower stock are left out of the analysis.
    3. Three weight fields — how much sell-through rate, stability and sales velocity contribute to the rating.
    4. Checkbox — when on, the “Can advertise” flag in the base catalog is updated from the rating.
    Demand analysis settings and their values
    FieldAllowed valuesDefaultWhat it does
    “Analysis period (days)”7…9030how many recent days of history points are used (how to change it)
    “History depth (days)”7…9030how many days of history are kept; the calculation sees nothing older than this, even if the period is longer
    “Min quantity for analysis”from 15items whose stock at the start of the “Analysis period” was lower don't get a rating
    “Weight: sell-through rate”0…10.5weight of “Sell-through %” in the rating
    “Weight: stability”0…10.3weight of “Stability” in the rating
    “Weight: sales velocity”0…10.2weight of average daily units sold
    “Keep inactive (days)”7…36530how many days to keep the history of a product that disappeared from the price list
    “Min rating”0…10050rating threshold for the “Advertise” mark
    “Min days of stock”0…902“Days” threshold for the “Advertise” mark
    “Min stability %”0…10030“Stability” threshold for the “Advertise” mark
    “Exclude trends”textcold,decliningtrends that block the mark: “Cold” and “Declining”
    “Update ‘Can advertise’ flag in product catalog”checkboxoncarries the mark over to the “Advertise” field of the product card

    The buttons at the bottom are “Save” and “Close”; after saving you'll see “Settings saved”. The next calculation picks up the new values: to see them in the grid right away, click “Recalculate”.

    The weights are normalized by their sum (how the rating is built). To turn a component off completely, set its weight to zero.

    There's no checkbox to turn on data collection in these settings: history collection is turned on only by the checkbox in each source card.

    What to do when a number looks odd

    Metrics that look odd and what to do about them
    What you seeWhyWhat to do
    The “Days” column shows “—” Nothing sold at all, or stock is already zero Check the “Quantity” column: zero means the supplier is out of the product (“Days”)
    “Growing” trend for a product that sells evenly In the second half of the period the price list arrived unevenly: the halves are compared by average daily units sold, and gaps in imports skew the picture Make decisions based on “Sold” and “Stability”, and treat the trend as a rough pointer (“Trend”)

    FAQ

    Why is there only one row in the “Metrics by period” table?

    One calculation is stored per product — for the current “Analysis period”. You can't show 7 and 30 days side by side (how to change the period).

    Why are the supplier's small items missing from the rating?

    A product whose stock at the start of the period is below the “Min quantity for analysis” (5 by default) doesn't get a rating, and there's no message about it. Lower the threshold in the settings.

    The product is no longer in the price list, but it's still in the window. Why?

    The nightly cleanup removes the history of disappeared products after “Keep inactive (days)”, 30 by default. Until then, the row stays with the numbers as of the last calculation.

    Can I compare the rating of products from different suppliers?

    Better not: in a price list with large stock, the same sales give a lower “Sell-through %” and therefore a lower score. The rating sorts products within one source (how the rating is built).

    Why doesn't any product have the “Advertise” megaphone?

    Most often the “Min stability %” threshold gets in the way: with infrequent imports, stability rarely reaches 30 %. Lower the threshold and click “Recalculate” (mark conditions).

    Where to find it and who has access

    Window
    “Supplier product demand analysis” — how to open it
    Access right
    the “Supplier product demand analysis” window in group permissions
    Prepare
    data collection turned on in the source and three imports on different days

    Related topics

    Was this article helpful?

    Missing a feature? Tell us.

    We will add it for free if other users need it too.

    If only you need it, we will agree on it separately.

    Article contents

    Suggest an improvement

    Try Elbuz free