Published Last updated Written and reviewed by the Resale Intelligence team
On this page
- What data does the valuation use?
- How is Expected Sell calculated?
- How is Safe Buy calculated?
- How does the system estimate how many days a device takes to sell?
- What are confidence and risk based on?
- How is true profit calculated?
- Does it use other shops' data?
- What are the limits of the estimates?
What data does the valuation use?
The system relies mainly on the prices your shop actually sold at. Listed prices, dealer prices and offers that your shop records are supporting data. Each source is weighted by how reliable it is, as the table below shows, and voided or returned sales and data older than 180 days are not used.
| Data type | Weight | Used for |
|---|---|---|
| Store's actual sold prices (after discounts) | 1.00 | Expected sell price |
| Store's buy prices | 0.70 | Showing purchase history |
| Recorded dealer prices | 0.50 | Dealer price |
| Recorded asking prices | 0.30 | Supporting data, adjusted down to the actual sold price level (×0.93) |
| Historical sales imported from files | 0.30 | Sold price and time to sell (the system did not see the sale happen, so it gets the same weight as manually recorded data) |
| Quoted/offered prices | 0.10 | Low-weight supporting data |
How is Expected Sell calculated?
Expected Sell is the weighted median of comparable prices, each adjusted to the device being valued for condition grade, battery, official or grey import and the age of the data. It is shown as a range spanning at least the 25th to 75th percentile, which widens when there is little data.
How prices are adjusted before calculating Expected Sell:
- Condition grade adjustment: A ×1.06, B+ ×1.03, B ×1.00, C ×0.88, D ×0.65 (each data point's grade compared with the unit's grade)
- Battery adjustment: below 80% ×0.94, 80–84% ×0.97
- Official/import adjustment: Thai official unit ×1.00, imported unit ×0.92, unknown ×0.97
- Time adjustment: value reduced by the age of the data, 2% per month by default
- Outliers removed using the IQR method when there are 5 or more data points
If there is not enough data overall (less than the equivalent of 1 actual sale or 3 asking prices), the system shows no price and says that there is not enough data yet.
How is Safe Buy calculated?
Safe Buy = Expected Sell − expected repairs − selling fees − holding cost − risk reserve − minimum profit, rounded down to the nearest hundred baht. The result is the highest buy price that still leaves the profit your shop has set.
- Holding cost = selling price × monthly rate (default 1.5%) × expected days to sell ÷ 30
- Risk reserve: 3% of the selling price by default
- Minimum profit: 10% of the selling price by default, and no less than ฿500
- The result is rounded down to the nearest hundred. Stores can adjust every value themselves.
What Safe Buy, Holding cost and Risk reserve mean is in the glossary; a worked example, line by line, is on the valuation page.
How does the system estimate how many days a device takes to sell?
It uses the median number of days the same model and spec took to sell in your shop. If there is not enough data it falls back to the model, then to your shop's default, and always says which one it used. This figure drives holding cost and the Price Curve.
Days-to-Sell is the number of days from when a device enters sellable stock until it sells. Price Curve shows 5 price points from 94% to 106% of the expected price, using your shop's real data once there are at least 6 sales; otherwise it is labelled an "estimate".
What are confidence and risk based on?
Confidence (0–100%) comes from the weighted amount of data, how spread out prices are, the share of data from the last 30 days and whether the data matches the exact spec. Risk has four levels, low, medium, high and very high, each shown with the reasons behind it.
- Confidence (0–100%) — Based on the amount of data (weighted), how spread out the prices are, the share of data from the last 30 days, and whether the data matches the exact spec or has to fall back to model-level data
- Liquidity (0–100) — Based on units sold in 90 days, speed of sale, sell-through, accumulated stock and price volatility
- Risk — Low / medium / high / very high, with reasons such as a price close to the selling price, above Safe Buy, condition, battery, volatility, slow sales, duplicate stock, an older model, or IMEI signals
What Confidence and Liquidity mean is in the glossary.
How is true profit calculated?
True profit = revenue actually received after discounts − every cost of the device: purchase price, repairs, battery, parts, shipping, fees, promotions, commission and storage. Every cost is recorded in a per-device ledger that cannot be changed after the fact.
Margin = True profit ÷ revenue · ROI = true profit ÷ total cost · Profit per day = true profit ÷ days held
Does it use other shops' data?
Today the system uses only your store's data. Province- or region-level data will be used only when there are at least 10 stores and 30 verified transactions in the area. It will be shown as anonymous aggregates, always keeping asking prices and actual sold prices separate.
What are the limits of the estimates?
Estimates come from past data, and the market can move faster than the data does. Shops with little data see low confidence or no price at all. The buying decision always stays with your shop: the system is a decision-support tool, not a price guarantee.