How businesses assess algorithmic price coordination between retailers under section 1 of the Austrian Cartel Act, preserve technical evidence and separate the relevant roles.
, Mag. Bernhard Brandauer, Rechtsanwalt
When several retailers use the same pricing software and change their prices in parallel, a central competition law question arises: do the prices reflect independent market reactions or coordination between competitors? Section 1 of the Austrian Cartel Act prohibits agreements and concerted practices that restrict competition by object or effect. It expressly covers direct or indirect fixing of purchase or selling prices.
A shared algorithm and a matching result do not by themselves prove prohibited coordination. The relevant questions are how the pricing instruction was created, which data were used, how retailers communicated and whether they could still set their own prices independently. Where trade between Member States may be affected, Article 101 TFEU must also be considered.
This article separates parallel prices from coordinated pricing, explains the technical evidence chain and distinguishes the roles of retailers, platforms and software providers. Personalised prices for customers and resale price maintenance imposed by a supplier raise different questions and are separated at the end.
This short path separates preparation, active use and a challenge. You can then send the relevant records to the firm.
Pricing rules, data sources, retailer roles and records determine the next review step.
This path provides initial orientation and does not replace an individual legal assessment.
Assemble the software description, provider material, data sources, parameters, test cases, approvals and internal responsibilities. Record which information comes from competitors and how each retailer’s pricing decision remains independent.
Preserve rule versions, configurations, price histories, data imports, change logs, messages and the actual price views. Assign the settings used and approvals given to each retailer.
Keep the complete letter and the challenged system version. Add logs, price lists, retailer communications, provider records, approvals and a chronology of changes.
Section 1(1) of the Austrian Cartel Act covers three forms of coordinated conduct: agreements between undertakings, decisions by associations of undertakings and concerted practices. The prohibition is triggered by an anticompetitive object or effect. Section 1(2) specifically includes direct or indirect fixing of purchase or selling prices.
A pricing algorithm is therefore neither a general safe harbour nor a separate prohibition in itself. The review asks who set or adopted the pricing rule, which information was available and whether retailers could still make independent decisions. An automated change may reflect an independent market response or execute a rule that was already coordinated.
Section 1(4) also treats certain recommendations to follow prices, price limits, calculation rules, margins or discounts as a cartel where they restrict competition. The technical format of a recommendation does not change the need to examine its purpose and implementation.
Matching price changes show an outcome first. Retailers may respond to the same public costs, demand changes, supply conditions or market prices. That does not by itself establish an agreement or concerted practice.
The review must therefore examine how the outcome was reached. Relevant points include direct messages, meetings, industry contacts, shared signals, coordinated price windows and whether a retailer aligned its conduct with an expected response from another retailer. The more a price rule is imposed or secured outside an independent market decision, the more carefully its creation must be documented.
A useful comparison covers several retailers, products and dates. Individual matching prices may be accidental or explained by a common market condition. A sequence connecting information, a price change and a response can provide a concrete line of inquiry. The assessment requires the complete data and communication chain.
The article on no-poach agreements between employers addresses a different horizontal arrangement. The common point is the review of conduct between competitors. The market and the substance of the arrangement remain different.
The technical creation of a price should be divided into separate steps for the evidence record. These include the source and update of data, the parameters used, the calculation, the output shown to a retailer and every later approval. A provider name or reference to artificial intelligence does not describe that process sufficiently.
Several retailers may use the same standard software and still act independently. The review risk can increase where a service provider combines competitively sensitive information from several retailers, sets shared targets or makes competitors’ price changes visible. These are factual questions and must not be inferred from use of the same product alone.
Preserve architecture overviews, data fields, interfaces, rule versions, user roles, configuration changes and test cases. The record should show whether a retailer can choose its own parameters, which data it can see and who approves a change. Different configurations and suggestions that were deliberately rejected belong in the file as well.
The businesses involved may perform different functions. A retailer may set prices, a platform may provide data and technical rules, and a service provider may configure or operate a model. For each participant, establish its own contribution, knowledge, influence and communications separately.
Central software does not automatically make several businesses cartel partners. Conversely, a formal description of independent pricing does not settle the issue where retailers adopt the same coordinated instructions or report deviations. The contract, technical implementation and actual conduct must be read together.
Build a chronology of the roles: who supplied which data, who changed the rule, who approved the price and who could influence the output? The article on multiple parties responsible for a competition violation concerns an unfair competition damages case. As a neighbouring documentation question, it illustrates why responsibility should be tied to specific contributions rather than company names alone.
A reliable file connects the visible outcome with the technical path that produced it. Preserve price views, timestamps, product and retailer details, incoming data, rule versions, system outputs, manual changes and approvals. Screenshots show the price, but they do not explain why it was displayed.
Keep messages, meeting records, tickets, provider communications, contract annexes, test cases and internal policies in the version that applied at the time. Exported logs should be linked to their creation time, the system used and the responsible person. Later corrections must remain identifiable as new versions.
The comparison should cover several retailers, regions, products and periods. Record prices that were not adopted and the reasons for manual interventions. This helps distinguish a binding algorithmic instruction from a recommendation or a technical calculation aid.
Before use, the business needs a clear description of its market role and pricing decision. Those responsible should define permitted data sources, exclude competitors’ sensitive information and document how retailers make their own pricing decisions. Approval should refer to a specific version of the rule and configuration.
The separation between generally available market information and sensitive information from another retailer is particularly important. A service provider should document which data enter a model, which outputs are shown to several customers and whether price recommendations or deviations are reported to other market participants.
Tests should cover different market conditions and deliberate deviations. Check whether retailers can reject suggestions or set their own prices and whether a deviation triggers technical, commercial or organisational consequences. A change to the data source, parameters, provider or distribution model requires a new approval.
The article on resale price maintenance concerns instructions from a supplier to a retailer. Algorithmic price coordination focuses on possible coordination among retailers at the same market level.
After a notice, the business should preserve the challenged state unchanged. This includes the specific price view, the rule in force, data imports, user roles, messages, approvals and the system version at the time. A later technical change does not automatically explain the earlier process.
Classify the challenge by sender, period, retailers involved, products and alleged coordination. Then separate what is documented from what is missing and from explanations based only on a technical assumption. External communications should describe the actual process precisely.
If rapid preservation or an injunction against the challenged practice appears necessary, the topic on injunctions and interim relief covers that separate protection path. This article addresses the preceding question of how pricing and coordination should be examined and evidenced.
Personalised prices require an assessment of whether individual consumers or consumer groups receive different prices based on automated decision-making. That concerns pricing towards customers and the related pre-contract information. The article on personalised prices addresses that customer-facing process separately.
Resale price maintenance concerns a retailer’s freedom to set its resale price against its supplier. Algorithmic price coordination concerns possible coordination among several retailers. One set of facts can touch both levels, but the assessment should keep them separate.
After a cartel has been established or alleged, a further question concerns affected parties, causation, loss and evidence. The article on cartel damages after price fixing addresses that claim path. This distinction prevents technical price matching from being treated as a completed damages case.
No. A matching price is an outcome that may result from the same market conditions or independent responses. The review must examine the information used, communications, pricing rules and the retailers’ actual freedom to decide.
No. Several retailers may use software independently. A closer review is needed where shared data, parameters, targets, notifications or economic pressure could influence independent pricing decisions.
The key issue is whether retailers can set their own prices and reject suggestions. Technical, commercial or organisational consequences of a deviation belong in the assessment. Calling an output a recommendation does not answer the question by itself.
Preserve price views, timestamps, data sources, rule and software versions, configurations, user roles, messages, approvals, test cases, provider records and change logs. Arrange the records by retailer, product and period.
Keep the challenged price, the system version at the time and the related communications unchanged. Add a chronology of data, rules, approvals and changes. The role of each participant and the alleged coordination can then be reviewed on a clear factual basis.
Organise specific contributions and responsibilities in a neighbouring case.
Review horizontal coordination in a different market and context.
Separate customer-facing pricing and automated decision-making.
Separate supplier instructions from coordination among retailers.
Assess affected parties, causation, loss and evidence in a claim.
Review evidence and possible urgent relief separately.
Receive new articles and legal guidance from the firm.
Topics
Give us a concise outline of the situation. We will clarify which documents are needed and which next step is appropriate.
Office
BRANDAUER Rechtsanwälte GmbH Giselakai 51 5020 Salzburg Österreich
Phone
+43 662 6280000