Analyze

Description & Overview

The “Analysis” page is intended for building predictive models and for conducting a detailed examination of the factors that influence the formation of demand forecasts.

Open

The page interface consists of two main sections:

• Parameters — used to select and configure the key forecasting settings.

• Results — displays the outcomes of the generated forecast and provides the ability to analyze the factors that influenced its formation.

Open

How MySales Builds a Forecast

MySales uses a high-performance mathematical forecasting engine, validated on hundreds of millions of “product–store” combinations.
The system automatically generates consistently high-quality forecasts for each item in every retail location, taking into account historical data and influencing factors.

1. Forecasting Hierarchy

The forecast is generated using a hierarchical approach — from more aggregated levels to more granular ones:

• Product category — entire network

• Product category — region

• Product category — store

• SKU — entire network

• SKU — region

• SKU — store

This approach makes it possible to leverage demand patterns at different scales and ensures forecast stability even when historical sales data is limited.

2. Data Used for Forecasting

With each recalculation, the system analyzes up to 3–4 years of historical data, including:

• actual sales

• prices and promotional activities

• inventory levels

• product hierarchy and reference data

• geographical structure of the network

• external factors (weather conditions, macroeconomic indicators, competitive environment)

Depending on the scale of the network, the volume of processed data can range from gigabytes to terabytes.

3. Overall Forecasting Logic

The forecast generation process is automatic, multi-step, and iterative.
The main stages include:

3.1. Data Preparation and Processing

The system loads historical data from databases and data warehouses, performs validation, and normalizes it.
For new products or new stores, data from similar items is used for periods where no own historical data is available.

3.2. Sales Cleansing

Historical sales are adjusted to account for factors that distort baseline demand:

• promotional activities

• price changes

• seasonal peaks

• lost sales due to stockouts

This allows the system to determine the true level of demand.

3.3. Seasonality and Trend Calculation

The system determines:

1. the nature of seasonal fluctuations (additive or multiplicative)

2. long-term demand trends

3. average and median sales values

Seasonality of the average price at the product group level is also analyzed.

3.4. Analysis of Influencing Factors

The system evaluates the dependency of sales on key factors:

1. price levels

2. promotions and discounts

3. weather conditions

4. macroeconomic indicators

5. cross-product interactions

If dependencies cannot be determined at a detailed level, data from more aggregated forecasting levels is used.

3.5. Forecast Model Generation

The system automatically:

1. calculates correlation coefficients

2. tests various forecasting models

3. selects the model that provides the highest accuracy based on historical data

A baseline sales forecast is then generated.

3.6. Incorporation of Future Factors

The forecast is adjusted to account for:

• planned promotional activities

• expected price changes

• forecasted values of external factors

The promotional effect may be determined using neural network models or by analyzing historical data from similar campaigns.

3.7. Short-Term Forecast Refinement

In the short term, the system accounts for the most recent sales dynamics and automatically adjusts the forecast.

3.8. Safety Stock Calculation

Safety stock is determined based on:

1. forecast accuracy

2. level of seasonal demand

3. sales variability

3.9. Final Adjustments and Use of Results

The final forecast can be:

1. adjusted by the user

2. used to generate pricing recommendations

3. applied to evaluate the economic impact of management decisions

4. Special Forecasting Algorithms

The system provides dedicated mechanisms for:

1. new products without sales history

2. distribution expansion to new stores

In such cases, aggregated data and demand patterns from adjacent levels are used.

5. Configuration Flexibility

The set of factors considered in forecasting is defined by the user and can be adapted to align with the company’s business processes.

The standard set of factors is described in the article Standard Set of Predictors

Parameters

The “Parameters” area is intended for selecting and configuring the key settings for building the forecasting model. It consists of the following categories:

• Groups

• SKU

Open

Groups

The dropdown button in the “Groups” field opens a list of available product groups.

The list displays previously used groups with their number and name (for example: 146. Bakery products, 3. Champagne, 130. Cider).

The “Select…” option opens an additional window where you can choose another group from the full list.

After selection, the group is displayed in the “Groups” field and is used for further data processing.

SKU

The dropdown button in the “SKU” field opens a list of available products.

The list displays previously used SKUs with their item number and product name.

The “Select…” option opens an additional window where you can choose other SKUs from the full list.

At the top of the field, a list of previously selected SKUs may be displayed. The “All” checkbox allows you to quickly select all SKUs within the chosen group.

Last Sales date

Open

The "Last Sales date" parameter defines the last week of historical data used by the system to generate the forecast.

This parameter is also used to validate the accuracy of the generated forecast.

Data Selection

Data selection is performed using dropdown lists and additional options.

Open

• The dropdown button opens a list for selecting the required data. The list displays previously used values and also includes a “Select…” option for adding new ones.

• The “All” checkbox allows you to select all items within the corresponding category.

• The “Exclude” checkbox allows you to skip forecasting for individual SKUs and perform forecasting only at the group level.

Settings

Open

The “Settings” button opens additional forecast parameters. To hide them, click the button again.

The following parameters are available:

Stores

The "Stores" parameter allows selecting one or multiple stores for which the forecast will be generated.

The selection can be made:

• by individual stores;

• by groups of stores;

• or by the value "All", which means applying to all stores.

The selected stores determine the set of historical data used by the system during forecast generation.

The button with a downward arrow opens the "Select stores" window, where one or more stores can be chosen.

At the top of the window, there is a field "Enter text to search", which allows quickly finding a store by number or name.

Below, a list of available stores is displayed with the following columns:

• Number — store code;

• Description — store name.

The checkbox at the top of the table allows selecting or deselecting all stores, while the checkboxes next to each row allow selecting individual stores.

The All Stores row allows quickly selecting all stores.

The "Replace" checkbox defines how the selection is applied:

• enabled — the new selection replaces the previous one;

• disabled — stores are added to the already selected ones.

The "Select" button confirms the selection, and the "X" button closes the window without saving changes.

Promo Drafts

The Promo Drafts parameter determines whether draft (unapproved) promotions are included in the forecast.

Options:

• All — includes all promotions (including drafts);

• disabled — includes only active/approved promotions.

Profile

The “Profile” parameter defines the type of forecasting model:

• 52-weeks — builds a forecast for 52 weeks

• 52-weeks log — builds a 52-week forecast with a detailed model-building log (available in the “Models” tab)

• 26-weeks — builds a forecast for 26 weeks

• Brain 1000 — builds a forecast using a neural network with 1000 iterations

• Brain 3000 — builds a forecast using a neural network with 3000 iterations

Supplier

The “Supplier” parameter allows selecting a supplier. The selection process is similar to selecting stores or groups.

Update

Clicking the “Update” button starts the forecast generation process.

Starting the Forecast Run

Clicking the “Run” button starts the forecast generation process.

During the calculation:

• information about the current forecasting stage is displayed under the progress indicator;

• after the calculation is completed, the system provides an audible notification.

Results

The “Results” window is automatically displayed upon completion of the forecasting process.
It is intended for reviewing sales forecast results, analyzing historical data, and examining the factors that influence the forecast.

The window consists of two main sections:

• the results display area

• the analysis area – intended for in-depth examination of the forecasting model, predictors, and their impact on the forecast.

Open

Open

Display Settings Panel

It is possible to configure the display of results as follows:

Open

• changes the display of the forecast date to weekly (Monday-based) or monthly;

Open

• allows selecting the forecast display range (from the start date to the end date) week in ISO format;

Open

• changes the display of the sales forecast by quantity or by sales value (in monetary terms).

To apply the changes to the parameters listed above, click “Show”

Additional parameters include a file management panel:

Open

The panel contains tools for working with files:

• Select file… — allows you to load a previously saved forecasting model for further analysis;

• Upload button — imports the model into the system;

• Save button — saves the built forecasting model in JSON format;

• Export button — exports the forecast results and model parameters in CSV format.

Analysis Area

The analysis area is represented by a header and an elements panel. It is an important part of the “Results” window and is used for detailed examination of the forecasting model.

Open

The header displays:

• the name of the product group or SKU;

• the number of stores or the store name;

• a button for exporting data in CSV format.

The elements panel provides detailed information on forecasting parameters. It consists of the following items:

• SKU;

• Stores;

• Components;

• Forecast;

• Predictors;

• Price;

• Trends;

• Models.

The main part of the analysis area is represented by a table with a weekly breakdown of indicators, which displays:

• actual sales;

• forecast values;

• previous sales for comparative analysis.

SKUs 

Open

The dropdown list allows selecting an SKU for analysis of the forecasting model.

After selection:

• the forecasting model for the selected SKU is displayed in a separate analysis area;

• the selected SKU is removed from the general list.

SKUs Chart

Area chart on SKUs

Open

The SKU decomposition chart is intended for analyzing the structure of product sales and the influence of various factors on demand formation. This tool allows you to view sales dynamics over time and evaluate the contribution of individual components of the forecasting model.

The chart can be configured using the Category and Component parameters, enabling analysis of different types of data and factors.

Category defines the type of indicators displayed on the chart. Available options include:

• Volume — displays the number of items sold;

• Value — shows sales in monetary terms;

• Original — displays historical sales values;

• Regressed — demonstrates factors influencing demand formation;

• Economy — shows the economic impact of various factors on overall sales performance.

The Component parameter allows selecting a specific indicator or factor for analysis. Available components include:

• Sales — actual sales volume for a given period.

• Forecast — projected sales values for future periods.

• Last Year Sales — sales volume for the corresponding period in the previous year (used for comparison).

• Order — volume of ordered goods (not necessarily sold yet).

• Seasonality — recurring seasonal fluctuations in demand (e.g., holiday peaks).

• Trend — long-term direction of sales (growth, decline, or stability).

• Reg. a.price (Regular Average Price) — standard product price without discounts.

• Reg. $ rate, % — percentage change in the regular price.

• Reg. temp. — absolute value of a temporary discount.

• Reg. temp. Δ — percentage value of the discount.

• Promo fcts (Promotional Factors) — factors related to promotions affecting price or demand.

• Min — minimum forecasted sales value.

• Auto mult. (Automatic Multiplier) — automatically calculated factor impact multiplier.

• Base — baseline demand level excluding marketing and external factors.

The chart displays the dynamics of indicators over time. Each component or SKU is represented by a separate color, allowing evaluation of its contribution to the overall result.

The X-axis represents time (weeks), while the Y-axis shows the number of units sold.

Data display can be controlled: clicking on the name or color indicator of an SKU in the legend temporarily hides the corresponding series from the chart. Clicking again restores its visibility.

This functionality enables more detailed analysis of individual product sales and simplifies comparison between them.

Using the chart, you can analyze:

• changes in sales across different periods;

• seasonal fluctuations in demand;

• the impact of price changes and discounts;

• differences between actual and forecasted sales;

• the role of individual products or factors in shaping overall demand.

SKUs Table

Summary table for SKUs

Open

The table is предназначена for viewing and analyzing the values of the selected component of the forecasting model in the form of a tabular report by products over time.

The tool allows comparing indicators across SKUs, time periods, and stores, as well as evaluating demand dynamics and the impact of factors on sales.

The first column contains a list of SKUs (specific products).

Across the top, time intervals are displayed in a weekly format (ISO week), allowing tracking of changes in indicators over time. Each cell of the table shows the quantitative value of the sales forecast for a given product in a specific week.

Components

Defines the model indicator displayed in the table.

Components are the indicators used for demand forecasting, order planning, and analysis of influencing factors.

1. Volume - calculation in units (Qty) — used for inventory planning.

• Sales — actual sales volume.

• Forecast — calculated demand for future periods.

• Last Year Sales — baseline for seasonal comparison with the previous year.

• Order — supply volume (ordered or planned replenishment).

• Seasonality — recurring seasonal fluctuations in demand.

• Trend — long-term changes in demand (growth, decline, or stability).

• Reg. a.price (Regular Average Price) — standard product price without discounts.

• Reg. $ rate, % — impact of promotions on price (percentage change).

• Reg. temp. Δ — impact of exchange rate (percentage change).

• Promo fcts (Promotional Factors) — relative change driven by exchange rate or promo-related effects.

• Min — minimum batch size or minimum order quantity.

• Auto mult. (Automatic Multiplier) — required order multiple (e.g., pack size or logistics constraint).

• Base — baseline demand excluding external factors.

• Anomaly / Lost — lost sales due to out-of-stock (OOS) situations.

• Excluded — data points excluded from model calculations.

2. Value - calculation in monetary terms (Value) — used for revenue planning.

• Sales — actual sales value for the period.

• Forecast — expected sales value calculated by the model.

• Last Year Sales — revenue for the corresponding period of the previous year, used for seasonal comparison.

• Order — value of generated or planned product supplies.

• Seasonality — impact of seasonal demand fluctuations on revenue.

• Trend — long-term sales trend.

• Reg. a.price (Regular Average Price) — standard product price without discounts.

• Reg. $ rate, % — impact of promotions on price (percentage change).

• Reg. temp. — temporary price adjustment (e.g., discount or markup) in absolute terms.

• Reg. temp. Δ — impact of exchange rate (percentage change).

• Promo fcts (Promotional Factors) — relative change driven by exchange rate or promo-related effects.

• Min — minimum batch size or minimum order quantity.

• Auto mult. (Automatic Multiplier) — required order multiple (e.g., pack size or logistics constraint).

• Base — baseline demand excluding external factors.

3. Original — used for model training.

• Seasonality — recurring cyclical changes in demand within a year.

• Trend — long-term direction of demand over time.

• A. price (Actual Price) — actual selling price in historical periods (after discounts).

• Price (Base Price) — standard listed product price before discounts.

• Price, % — percentage change in price relative to a baseline.

• Discount — absolute value of the price reduction.

• Discount, % — percentage reduction from the base price.

• Stock — available inventory level.

• Item count — number of units sold or available.

• $ Rate — exchange rate used for price or cost conversion.

• $ Rate, % — percentage change in the exchange rate.

• Seas. mult (Seasonality Multiplier) — coefficient representing the strength of seasonal impact on demand.

• Avail (Availability) — proportion of time the product is in stock and available for sale.

• Seas. type (Seasonality Type) — category describing the product’s seasonal demand pattern.

• Temperature — observed temperature affecting demand.

• Temperature, Δ — change in temperature compared to a reference period.

• Rain — level of rainfall affecting demand.

• Snow — level of snowfall affecting demand.

• Promo ratio — share of sales occurring under promotional conditions.

• Promo ID — identifier of a specific promotion.

• Promo XL — indicator of high-impact or large-scale promotions.

• Infl. price — price adjusted for external factors (e.g., inflation or normalization).

• Infl. disc — discount adjusted for external factors.

• Promo uplift — increase in demand attributable to promotional activity.

• Min — minimum order quantity constraint.

• Auto mult. (Automatic Multiplier) — required order multiple (e.g., pack size constraint).

• Median — median value representing typical demand.

• Base — baseline demand excluding external influences.

4. Regressed (influencing factors) — used to explain demand.

• Regressed (Influencing factors) — variables used by the model to explain and quantify the impact of different drivers on demand.

• Reg. a.price (Actual Price) — effect of the actual historical selling price on demand.

• Reg. price (Base Price) — effect of the base (listed) price on demand.

• Reg. price, % — effect of relative price changes on demand.

• Reg. disc. (Discount) — effect of the absolute discount value on demand.

• Reg. disc., % — effect of the relative discount level on demand.

• Reg. stock — effect of available inventory on demand.

• Reg. $ rate — effect of the exchange rate level on demand.

• Reg. $ rate, % — effect of exchange rate changes on demand.

• Reg. count — effect of the number of units (or items) on demand.

• Reg. temp. (Temperature) — effect of temperature on demand.

• Reg. temp. Δ — effect of temperature changes on demand.

• Reg. rain — effect of rainfall on demand.

• Reg. snow — effect of snowfall on demand.