Skip to content
Procurement management

Predictive analytics
for Kaspi
sellers

Forecast demand, avoid stockouts, and optimize purchasing with AI

87%
Accuracy
−31%
Overstock
−84%
Stockouts

90-day forecast

Why sellers lose money on procurement

Problem

Purchasing by guesswork

Sellers buy by guesswork and lose money. Overstock ties up capital, while stockouts mean missed sales and lost search ranking.

Capital is tied up in excess inventory
Stockouts drop your search rankings
Competitors capture your customers
AWW Solution

Accurate forecasts instead of guesswork

AWW predictive analytics analyzes sales history, seasonality, and trends — forecasting demand for 30–90 days. You buy exactly what you'll sell.

90-day forecast for every SKU
Alerts before stock runs out
Detects trends 2–3 weeks earlier
Features

Everything for accurate purchasing

30–90 day demand forecasting

LSTM and Prophet algorithms analyze history and external factors to predict sales volume for each SKU.

Stockout alerts

The system warns in advance which items will run out in a week. You can reorder before losing sales.

Trend detection

The model spots category growth 2–3 weeks before the peak. You buy earlier than competitors and capture the gains.

Inventory optimization

Accurate reorder recommendations for each item. Minimum tied-up capital, maximum turnover.

A Kaspi seller reviewing product profit analytics in the AWW app
Monthly profit +18 %
Top product found

Profit per SKU

Numbers you can base decisions on

AWW collects your store's orders and calculates net profit — by day, week, month, and for each item individually. The repricer uses the same cost-price data: it won't drop a price below the break-even point even at the height of a price war. And the economics of a new purchase batch are easy to estimate in advance with the profit calculator.

Download the app
How it works

Three steps to precise purchasing

1

Connect your data

Integrate your Kaspi store. The system will automatically import sales history and current inventory.

2

Get forecasts

The algorithm will analyze the data and provide demand forecasts for 30, 60, and 90 days for each product.

3

Make precise purchases

Follow purchasing recommendations. Receive alerts about shortages and trends in real time.

Results

Numbers, not promises

87%

demand forecast accuracy

−31%

reduction in overstock

−84%

reduction in product shortages

Profit analytics, not turnover

Three things that change how you see your own store

Kaspi store sales growth: AWW profit analytics
Net profit · per SKU

Why revenue is deceptive

Turnover can grow month after month while the money in your account doesn't. Between the selling price and what you actually earn stands a chain of costs: the Kaspi commission with its own rate for every category, delivery, packaging, returns, and the cost of the goods. If you only look at total sales, those costs stay invisible — and a store can trade at a loss for weeks without suspecting it.

AWW counts net profit per SKU, not turnover: it collects the store's orders, subtracts the category commission, delivery, and cost price — and shows how much you actually earned on each item per day, week, or month.

Kaspi store order tracking: every product under analytics control
Product margin · −2 %

Products that churn turnover for nothing

The biggest discovery for most sellers is items with plenty of orders and near-zero or negative margin. They create a feeling of brisk trade, tie up money in purchasing and time in handling, but bring no profit.

Sorting your assortment by net profit surfaces such products in a minute. From there the decision is yours: raise the price, switch suppliers, or drop the item from the lineup. Before buying a new batch, it's worth checking the economics in the profit calculator — it accounts for Kaspi commissions across 21 categories.

AWW push notifications about competitor prices on Kaspi
Competitor lowered a price
Competitor price history

Price history and alerts at the right moment

For every product you can see how competitors' prices changed and when a competitor started dumping — so it's clear where a margin dip came from. Together with the repricer it forms a closed loop: analytics shows where profit leaks, and the repricer automatically keeps the price within your set bounds.

Push alerts arrive when something important happens: a competitor changed a price, a product lost the Buybox, or an item's margin went negative. No need to keep the dashboard open all day — your phone will tell you when it's time to step in. To see your own store's profit, download the app: the first numbers appear right after order sync.

Real results

Data, not guesswork

Sellers on Kaspi.kz make decisions based on AWW predictive analytics

Stop buying blindly

Enable predictive analytics and make data-driven decisions

Try for free
FAQ

Frequently Asked Questions

For a basic forecast, 30 days of sales history is sufficient. The more data, the more accurate the forecast — ideally at least 3 months. The system starts working immediately and improves accuracy as more data accumulates.
The model analyzes sales history, current inventory levels, prices, seasonal patterns, holidays (Nauryz, March 8, Black Friday) and category trends on Kaspi.kz.
The algorithm automatically detects seasonal patterns in your data and adjusts forecasts accordingly. Kazakhstan's holidays and peak periods are already incorporated into the model.
For new products without sales history, the system uses data from similar items in the category and overall market trends. Accuracy increases as your own data accumulates.
Forecasts are recalculated daily with new sales data. Alerts about shortages and anomalies are delivered in real time.
The system provides precise purchase quantity recommendations for each SKU. You can see how many units to order and by what date to avoid stockouts.