Incremental

A one-sentence summary

Incremental is a causal intelligence platform for retail and business media, designed for large brands and agencies. It aims to determine exactly how much additional sales advertising has generated, rather than attributing all transactions that were reached through advertising to the efforts of the media.

What is incremental?

Incremental is operated by Tradeswell, Inc. under the Incremental brand; it focuses on providing marketing metrics, optimization, budget planning, and sales forecasting across different retailers and media channels. The product offers daily data on incremental sales and iROI at the level of individual campaigns and SKUs, rather than only providing quarterly attribution reports.

It is not the same product as INCRMNTAL, which has a similar name and uses a spelling that omits vowels. Research, procurement, and SEO descriptions should be based on Incremental’s current commercial media causal measurement framework.

Why not just look at ROAS?

Traditional ROAS attributes all sales that occur after the ad is shown to the ad itself, without taking into account those who would have purchased anyway, as well as factors such as promotions, inventory levels, search rankings, and seasonal influences. An incremental measure, on the other hand, focuses on the difference between the situation with the ad and the situation without it, thus providing a more accurate reflection of the actual economic impact.

Main functions

Causal measurement across retailers and channels

The platform transforms various retail media networks, channels, and campaigns into a unified causal measurement framework. Brands can compare the incremental sales generated by the same investment across different retailers and campaigns, without being affected by the proprietary attribution methods used by each platform.

iROI at the activity and SKU level

The results can be detailed down to individual campaigns, ad groups, and SKUs, or aggregated upwards to channels and brands. The team can identify campaigns with a high ROAS but low growth, as well as those that appear to yield average returns but actually generate new demand.

Commerce Graph business graph

A business map connects products, campaigns, promotions, search rankings, prices, inventory, the Buy Box, and competitive dynamics. As a result, the model is able to understand the retail environment in which advertising takes place, rather than focusing only on impressions, clicks, and orders.

Multi-framework causal model

By combining incremental econometric methods, design experiments, and synthetic experiments, it is possible to estimate the media impact of different retailers, channels, and SKU’s. The use of various inference frameworks helps with cross-validation, but it cannot eliminate all the uncertainty associated with counterfactual estimates.

Daily learning cycle

The model compares the short-term predictions with the actual results on a daily basis, and uses the differences to update subsequent estimates. The system gives priority to recent data, enabling changes in activity levels, promotions, inventory levels, and competitive conditions to be taken into account more quickly.

Optimization suggestions

The system generates budgeting and bidding adjustment suggestions based on iROI, helping teams allocate funds to activities that are more likely to generate additional sales. These suggestions can be integrated into the existing media purchasing workflow, reducing the need for manual re-entry of data after analysis is completed.

Scenario planning and sales forecasting

Users can simulate the level of media investment required to meet sales targets, or compare the impact of shifting budgets from one channel or campaign to another. These predictions are useful for guiding discussions on planning strategies, but they should not be regarded as financial commitments that guarantee achievement of goals.

Coordination between external media and retail

The platform analyzes not only the retail media available on its own site, but also takes into account the impact of external media on product discovery, brand awareness, and the performance of activities on the site. This allows it to identify the halo effect that one channel can have on other channels or retailers.

Calibration of experimental results

Incremental does not carry out A/B tests on its own, but it can receive the results of experiments conducted by platforms such as Amazon and Walmart, which can then be used for initializing, validating, and calibrating models. Combining experiments with models makes it easier to identify deviations compared to relying on just one method alone.

How does the system work?

Data input

Results and delivery methods

The process from access to optimization

1. Define the scope of retailers, channels, brands, campaigns, and SKUs to be measured.
2. Inventory media, sales, shelves, prices, promotions, stock, and experimental data.
3. Authorize the platform through existing connectors, or define the method for delivering customer data.
4. Standardize products, promotions, retailers, dates, currencies, and metrics.
5. Check for data gaps, outliers, duplicate records, and cross-platform mapping.
6. Have the platform create a business map and the initial set of causal models.
7. Calibrate the model using historical experiments, business events, and financial results.
8. View activity and incremental sales at the SKU level, along with confidence intervals and iROI.
9. Review budget proposals and submit them to the advertising platform within defined limits.
10. Continuously compare recommendations, executions, forecasts, and actual sales to gradually increase automation.

Verification process before procurement

1. Select a set of activities that have stable historical data and explainable business events.
2. A portion of the data is kept as a blind test and is not used during the model development process.
3. Compare the results with existing Holdout experiments, retail platform tests, and financial data.
4. Check whether the model handles shortages, promotions, price changes, and Buy Box variations correctly.
5. Evaluate the error rates for activities, SKUs, and cross-retailer mapping.
6. An explanation is required for causal methods, assumptions, confidence intervals, and failure conditions.
7. Test optimization suggestions using a small budget; do not implement them fully automatically right away.
8. Confirm data permissions, retention, deletion, cross-border transfers, and arrangements for sub-processors.
9. Include accuracy, update frequency, as well as interface and service support in the contract.

Connection and activation

The integration of new e-commerce or media platforms usually takes around 2 to 4 weeks, but the complexity is influenced by factors such as authentication requirements, historical data, the quality of the fields, and the limitations of the platform’s interfaces. Pre-existing connections also require that the region, account type, and specific datasets be specified in the contract.

Pricing and procurement methods

The price information was verified on August 23, 2026; the actual amounts, taxes, exchange rates, and discounts may vary, and the final figures will be those displayed on the settlement page.

Incremental does not offer any fixed packages, monthly or annual fees, nor a free version; purchases are typically made through scheduled demonstrations and direct sales to businesses. The budget required is determined by taking into account the number of retailers and distribution channels, the scale of events and SKUs, historical data, connectors, the range of models, as well as service support needs.

Is there a free trial available?

There is no free version or standard trial that can be registered directly. Companies usually need to schedule a demonstration first, and then discuss matters related to data access, the scope of the pilot project, and the commercial terms.

Which users are it suitable for

• Large consumer brands that are advertised across multiple retail media networks.
• E-commerce teams that need to standardize the metrics used for Amazon, Walmart, and other sales channels.
• Advertising agencies that are responsible for planning, purchasing, and optimizing retail media campaigns.
• We hope to upgrade ROAS to a marketing analysis team that focuses on incremental sales and iROI.
• Operational teams that need daily results at the activity and SKU level, rather than quarterly reports.
• A business analysis department that has data on promotions, inventory, shelf space, and competition.
• I hope to refer these causal recommendations back to the teams at Skai, Pacvue, Flywheel, or WPP Open.
• Brands that have sufficient historical data and experimental results available for model validation.

Situations that are not very suitable

• Individual creators and businesses with limited advertising budgets.
• Users who only need to perform simple clicks, complete conversions, and view the platform’s ROAS reports.
• Brands that lack SKU-level sales, media, promotion, and inventory data.
• Teams that wish to register immediately and use the service at a low, public price.
• Projects that feature only a single short-term campaign and lack a historical baseline.
• Companies that require tools to directly replace the advertising platforms and retailer backends.
• Decision-makers who cannot accept model estimates, uncertainties, and ongoing calibration.
• Teams that need to make their developer APIs, SDKs, or self-hosted open-source versions available to the public.

Main advantages

• It focuses specifically on retail and commercial media, rather than general digital advertising attribution.
• Place media, sales, pricing, promotions, inventory, and shelf signals in the same diagram.
• Economic impacts are estimated using various causal frameworks, not relying solely on the final touchpoint.
• Models and results are updated daily, making it suitable for frequent optimization.
• Supports iROI at the activity, campaign, and SKU levels.
• It also provides measurement, recommendations, scenario planning, and sales forecasting.
• The suggestions can be integrated into existing media purchasing platforms.
• It does not sell ads, operating independently of retailers, media sellers, and bidding platforms.
• The current solution emphasizes the elimination of the need for implanted tags as well as direct identity information.

Usage restrictions and risks

• Causal models estimate counterfactuals; they do not equate to directly observing another reality.
• Daily training does not automatically correct missing, inaccurate, or incorrectly mapped data.
• Results at the activity and SKU levels may be unstable due to insufficient samples.
• Bias can occur when promotions, stockouts, competition, and external media are not fully incorporated into the model.
• The platform does not carry out A/B tests directly; instead, retailers or other tools are needed to conduct such experiments.
• An overall increase in a manufacturer’s sales does not guarantee that individual brands will achieve the same results.
• There is no fixed price on the public pages, and transparency regarding procurement and budgets is limited.
• Automatic suggestion delivery involves actual budgets; approval processes, limits, and rollback options should be maintained.
• The technical details of the complete model, the criteria for determining confidence intervals, and the full list of connections are not made public.
• The privacy policy updates are made publicly available at an early stage; it is therefore necessary to verify the current data arrangements during the contract phase.

Privacy and data processing

The current solution emphasizes that no direct identity information is required, nor is it necessary to insert tags on the brand’s website. The platform obtains business data through interfaces provided by retail, media, and e-commerce platforms, and then processes and standardizes this data on a daily basis.

However, the 2023 version of the privacy policy states that customer service data may include demographic information, purchase history, Cookies or device identifiers, as well as hashed email addresses and mailing addresses. Companies should distinguish between current causal measurement methods and historical data management systems, and determine whether any identifiable or linkable data is being processed in a given project.

Privacy check during procurement

• Obtain the latest DPA, list of data fields, and list of sub-processors.
• Verify whether the current solution accepts Cookie IDs, device IDs, or hashed contact information.
• Distinguish between anonymous, de-identified, pseudonymized personal information and personal information in a legal sense.
• Define the retention periods for raw data, normalized data, model parameters, and outputs.
• Verify whether the data is used for cross-client modeling, product improvement, or benchmark analysis.
• Clarification is requested regarding U.S. custody, cross-border transfer, and deletion verification.
• Review the permissions required to activate the deployment platform as well as the controls related to budget setting.
• In the case of an older privacy policy, the new contract and current security documents shall prevail.

Security and governance recommendations

• Use the minimum required permissions for connections to each retailer and distribution platform.
• Separate the access to read-only data from the permission to modify budgets.
• Retain model versions, data versions, recommendations, approval, and execution records.
• It is recommended to set upper limits for activity-level budgets and to implement mechanisms to detect abnormal changes.
• Re-authorize regularly and remove inactive brand, channel, and agent members.
• Request the current security audit, encryption, backup, and incident response documents before making a purchase.
• Set up alerts for interface interruptions, changes in retailer fields, and model drift.
• Major budget transfers are first subjected to limited pilot tests and financial reviews.

API, GitHub, and open-source status

• The platform connects retail, media, and e-commerce systems through APIs and other means.
• The results can also be uploaded to the advertising platforms and the customer’s internal data lake.
• No developer API documentation available for the public has been identified.
• No official public SDK, command-line tools, or Webhook directory were found.
• No official GitHub open-source repository for the current Incremental platform has been identified.
• There is no disclosure regarding self-hosted deployments, model weights, or source code licenses.
• The integration of interface-based products does not mean that the public can apply for API keys on their own.
• Incremental should be regarded as a proprietary enterprise cloud platform.

Basic information

Frequently Asked Questions

Are Incremental and INCRMNTAL the same company?

No. Although their names are similar, the products, companies, and websites involved are different; this entry is about Incremental, which is intended for retail and commercial media.

What is incremental sales?

Incremental sales refer to the additional sales that result from advertising, as compared to what would have happened in the absence of that advertising. It attempts to account for natural demand, promotions, inventory levels, prices, and other factors, rather than including all transactions that result from advertising efforts.

What is the difference between iROI and ROAS?

ROAS is typically calculated by dividing the sales attributable to advertising efforts by the cost of those ads, while iROI measures the return on investment based on the estimated incremental value. iROI is more stringent in its approach, but it also relies on causal models and data quality.

How often are the results updated?

The platform releases data and results on a daily basis, and updates the model by comparing short-term predictions with actual values each day. The specific timing of these updates and any associated delays depend on the individual retail and media interfaces.

Is it necessary to install tracking tags on the website?

The current solution does not require the use of website tags; data enters the system primarily through connections with retail, media, and e-commerce platforms, as well as through the provision of customer data.

Is personal identification information required?

The current marketing strategy emphasizes that direct identity information is not required, but earlier privacy policies allowed the processing of customer service data such as device identifiers and hashed contact details. For actual projects, a list of data fields and a contract confirmation are necessary.

Will it run A/B testing directly?

It will not run directly. The platform can utilize the experimental results already available from retailers and other channels to initiate, calibrate, and validate the model.

Can the advertising budget be adjusted automatically?

Optimization suggestions can be integrated into workflows such as Skai, Pacvue, Flywheel, and WPP Open. Companies should still implement manual approval processes, budget limits, mechanisms for detecting anomalies, and options for rolling back changes.

What is the price?

There are no fixed public prices or packages available. Brands and agents need to schedule a demonstration in order to obtain a quote, which is determined based on data volume, channels, campaigns, SKU, as well as the scope of connections and services offered.

Is there a free version available?

No free self-service version or standard free trial has been identified. Companies can discuss the terms for demonstrations or trials with sales.

Are API or open-source versions available?

The product features platform-level interface integration, but no public developer APIs, official SDKs, open-source code repositories, or self-hosted open-source versions have been identified.

Summary

Incremental is suitable for large brands and agencies that need to compare the actual incremental impact of media across different retailers, and that use daily performance data as well as iROI at the SKU level to make decision-making about advertising investments. Its key features include business mapping, multi-framework causal inference, as well as daily learning and optimization processes.

Before making purchases, the model should be validated using real historical data and experimental results; the overall cost in the absence of fixed prices must be determined, and any discrepancies between the old privacy policies and the current policies that do not include PII information need to be addressed. Any automated budget recommendations should be implemented under frameworks that provide approval controls, limits, and the ability to revert to previous settings.