Advertising Intermediation After the Xandr Sanction

1. The design behind an advertising intermediary’s revenue

On October 12, Norway’s Datatilsynet announced Xandr’s NOK 92 million sanction over invalid Grindr consent, covering 2018–2020 processing.^[1]

For advertising markets, the useful starting point is the transaction system an intermediary provides. Connecting sellers and buyers raises questions about which information reaches which recipients and supports transactions at which prices. Convenient defaults reduce onboarding work. Measuring revenue and costs after changing them nevertheless requires aligning participants, demand and volume. The analysis must move from a feature list to the conditions behind earnings.

The fine cannot be converted directly into a decline in global advertising revenue. However, a reassessment of the intermediary’s role could change the information transmitted, the buyers connected and the work required to check consent. Those changes could affect inventory prices and operating costs. The analytical path runs from legal role to transaction conditions, then revenue and costs, and finally business value. Starting and ending with the size of the fine skips the conditions in between.

SG Group’s view is that advertising intermediation should not be valued through one simple relationship in which more participants always mean more value. More participants and information can expand the choices offered to buyers. They can also expand the burden of explaining who is responsible for which processing. Both effects can arise from the same design. The durability of the business depends on the combination of transaction reach and the ability to carry the resulting responsibilities, rather than on the fine alone.

2. What does a service measure?

Under the EU’s General Data Protection Regulation (GDPR), controllers determine processing purposes and essential means. European Data Protection Board (EDPB) guidance includes information, duration and recipients among those means. Multiple participants are not automatically joint controllers, and processors can have implementation discretion.^[3]

Changes that speed up communication and changes to buyer connections reach prices through different channels. The former calls for examining completion rates and response times; the latter requires examining what added buyers purchased. Both may be called configuration changes, but the outcomes measured differ. Simply comparing revenue before and after would combine speed improvements with changes in demand. Defining the outcome first helps align comparison periods and scope.

Consider an illustrative company offering both inventory-selling support and purchasing support. The seller-side service measures inventory sold and money reaching publishers; the buyer-side service measures purchases and outcomes. Calling something one transaction does not make bid, purchase and delivery counts equal. Different denominators when dividing each service’s fees by volume can change comparisons of prices and costs.

When assessing a service with added features, connect the revenue category to the transaction units that produce it. The label may still say advertising technology, but new outcomes or charging conditions can make earlier units harder to compare. It is useful to establish whether rising counts reflect demand or a change in counting. Reading contracting entities and service scope in CFD-provider comparisons likewise offers a foundation for moving beyond names to what is provided.

Reading contracting entities and service conditions — Free guide to entities, jurisdictions and scope; not evidence determining Xandr’s role.

3. Connect setting changes with price comparisons

Despite customer exclusions, Xandr’s framework, defaults and recipient decisions mattered to the joint-controller finding for buyer-side and seller-side processing.^[2]

Higher average advertising prices after narrowing settings do not immediately establish that the change raised prices. Removing low-price transactions while retaining higher-price ones can raise the average even as total revenue declines. Comparing revenue requires matching inventory types, demand periods and quantities sold. Avoiding the conversion of average-price improvements into whole-transaction improvements helps identify what a change in defaults gained and sacrificed.

Economically, defaults reduce the work required to start using a service. Sellers can begin trading without building a separate connection to every buyer, while buyers can reach many sellers more easily. Precisely because that benefit exists, the standard configuration can materially affect transactions. The ability to change it later is different from the initial configuration having no effect. If broad defaults are a competitive advantage, explaining the processing that supports them becomes part of the service’s value.

It is also useful to separate improvements in communication from improvements in advertising prices. The IonQ–Elisity article’s communication permissions and encryption design offer a different example of reading technical capabilities. Better communications infrastructure does not necessarily increase advertiser demand. Conversely, higher advertising prices from stronger demand do not prove improved communication. Keeping indicators from different stages separate helps explain which investment produced an effect.

IonQ–Elisity: communication permissions and encryption — A separate communication-permissions case, distinct from determining lawful data use.

4. Information gains meaning through combinations

The authority assessed app use combined with advertising identifiers and IP addresses as sensitive data concerning sexual orientation, not a finding of an individual’s orientation.^[2]

In advertising transactions, the detail contained in a single field does not fully describe its effect. Consider, for illustration, information identifying a type of service and a separate number distinguishing a device. The first describes a service, but linking it to the second can turn it into information about a device’s use. Combining information may improve advertising relevance while also enabling inferences about users. The same combination that increases commercial value can widen the scope requiring protection.

Counting fields can therefore misstate the value of selling or processing information. A small number of linked fields may carry strong meaning; many separate aggregate figures may not support the same inference. Describing a technical identifier as something other than a name does not resolve that distinction. When a company proposes to raise prices by adding data, revenue analysis also needs to ask which combinations create value and under what conditions those combinations can be used.

More detailed information is not necessarily more useful to an advertiser. Even if it reaches an appropriate audience, the nature of the information may add checking and explanation costs, changing the total cost per outcome. Conversely, if the advertising objective can be achieved with less information, collecting more becomes less valuable. There remains room to compare the objective, necessary information and transaction conditions on a consistent basis.

5. Collecting consent and explaining processing are different jobs

The authority found consent invalid and Xandr’s verification of validity and scope insufficient. Contractual exclusions alone did not exempt it.^[2]

From a transaction-efficiency perspective, it makes sense for the service interacting with users to collect consent and for other participants to use the result. Separate explanations from every company could increase onboarding work and fragment the information presented. However, the value of that division of work depends on recipients being able to interpret the result. A brief “consent given” status may not align the information, purpose, recipient and time. If later stages have to repair those gaps, some of the original efficiency is lost.

EDPB consent guidance covers records demonstrating past explanations and workflows, potential invalidation of original consent through substantial processing changes, and avoiding unnecessary information collection.^[4]

Operating metrics also need denominators. Consider matching work at two publishers: organising many records at a small publisher does not provide equivalent coverage of the larger publisher’s transactions. Publisher count, record count and covered transaction volume are different measures. Judging progress solely by record counts can miss where revenue-supporting transactions improved. Matching the work’s scope with the volume it supports makes the use of operating expenses easier to compare.

6. SG Group analysis: connections and inventory value are different measures

The first analysis is that more connections do not necessarily increase an advertising intermediary’s profit. In an illustrative transaction, adding a buyer creates the possibility of another auction participant. That buyer may not participate every time, and participation may not improve clearing prices or sales volume. Connection count measures opportunities. Raising the value of inventory requires two further steps: participation and competition.

Suppose that the additional connection also creates costs for managing the recipient, matching the processing to consent and reviewing changes in conditions. Its economic value depends on whether incremental income exceeds those costs. Adding a connection is not inherently harmful. A buyer with distinctive demand may value inventory that existing buyers did not purchase. The distinction is that the number of recipients and the number of transactions that increased profit are different measures.

The hypothesis gains support if recipients can be reduced without changing fill rates or prices, while management costs decline. Conversely, the loss can be large if an excluded buyer has demand that other buyers cannot replace. Relevant indicators include the inventory the additional buyer actually purchased, its contribution to prices and the associated operating costs—not just total connections. The comparison must also stay within the conditions permitted for processing users’ information.

This approach separates making transaction counts look larger from increasing what sellers retain. If an intermediary chooses a broad recipient network, actual advertising demand offers a stronger explanation for that choice. A long menu does not itself establish quality; treating it as such converts feature quantity into economic performance. Following Xandr’s case, service assessment can usefully bring both the value of deciding recipients and the burden of that decision back to a common transaction unit.

7. What changes profit when compliance costs increase?

The second analysis separates fixed costs from costs incurred for each transaction. For illustration, let transaction volume be q, income per unit 5, variable cost per unit 1 and period fixed costs 4. Money uses common hypothetical units, and q does not represent real advertising impressions. Profit is 4q−4: at q values of 1, 2 and 4, it is respectively 0, 4 and 12. The equation makes the deductions from revenue explicit; it does not estimate Xandr’s results.

Suppose building a system to check transaction conditions raises only fixed costs, from 4 to 8. Profit becomes 4q−8, giving −4, 0 and 8 at the same volumes. Break-even volume rises from 1 to 2. This form of cost increase can explain why a larger business spreads fixed costs more widely. It depends, however, on being able to reuse the infrastructure. If each participant requires a separate response, that scale benefit becomes smaller.

In another case, fixed costs remain 4 while the variable cost of checking each transaction rises from 1 to 2. Profit becomes 3q−4, giving −1, 2 and 8 at q values of 1, 2 and 4. More volume does not eliminate the additional cost per unit. Both cost-increase cases happen to produce profit of 8 at q=4 under these assumptions; that does not make their cost structures equivalent. Their different results at smaller volumes show why the form of the response matters.

For a real company, the useful questions go beyond the label “compliance costs”: what is being rebuilt, and where does expense recur? Investment in shared infrastructure needs to be assessed for reuse across later transactions. Individual checks require attention to their scope and frequency. If changes in usable advertising information also change income per unit, a cost-only equation is insufficient. Connecting the financial statements helps distinguish expense recognition, investment and cash payments.

Connecting statements and financial notes — Free guide to earnings, resources, cash and notes; not observed financial data for this case.

8. When the burden falls on smaller publishers

Under the fixed-cost assumption, allocating the same sum across fewer transactions produces a greater unit burden. Dividing the earlier fixed cost of 4 by volumes of 1, 2 and 4 gives 4, 2 and 1 respectively. An intermediary that can share checking and explanation systems may lower the burden compared with a small publisher building the same capabilities alone. This also challenges the simple conclusion that examining responsibilities necessarily destroys the intermediary’s value.

Whether publishers receive the benefit of shared infrastructure depends on pricing. A provider may achieve fixed-cost savings but raise its fees, leaving small publishers no better off. Their bargaining power may also be limited if alternatives are few and migration requires rebuilding settings and connections. Evidence of scale economies does not automatically establish a benefit for service users. Fees and the capabilities provided need to be assessed together.

If an advertising-funded publisher cannot absorb additional costs, it might change advertising volume, user fees or available features. The choice depends on audience responses, other income and opportunities to reduce costs. Revising advertising-processing conditions does not inevitably raise subscription prices. Nor does the presence of a paid service establish that all data processing has stopped.

Global economic transmission would emerge by adding up such company-level changes. Publisher exits, paid services or greater concentration among shared-infrastructure providers could alter information and advertising distribution. This single case cannot measure that spread. Assessing international impact requires further evidence about which processing faces the same issue, how many publishers remain viable after responding and where advertisers move their spending.

9. Between advertiser spending and publisher receipts

The third analysis concerns the adjustment between what advertisers pay and what publishers receive. Even with unchanged advertiser spending, checking costs passed through an intermediary’s fees could reduce publishers’ receipts. Alternatively, advertisers could pay additional fees while publishers retain their income. The place where a cost arises is not necessarily the place that ultimately bears it. Contract terms, alternatives and bargaining power determine its transmission.

Advertisers’ assessment of cost per outcome also matters. If less information produces fewer outcomes for the same spending, they may reconsider budget allocation. Yet protecting user relationships or a brand’s reputation may also have value, so short-term outcomes need not be the sole criterion. Placing the cost of achieving the objective and the cost of maintaining processing conditions in the same evaluation period explains demand shifts better than reading advertising spending through one indicator.

If advertisers move spending to another publisher or method, one publisher’s loss can coexist with another’s gain. That is not an equal contraction of the entire advertising market. Separate cases include lower total spending, unchanged spending with a different allocation, or higher spending to cover added costs. To connect transaction changes with equities, begin by identifying which companies receive demand and which absorb costs, rather than assigning one direction to the entire industry.

Financial-advertising review examines what an advertisement communicates and the evidence behind its claims. This case concerns information processing before that advertisement reaches its audience. Both stages need explanation, but completing one review does not establish the suitability of the other. Advertisers comparing total costs can better identify improvements by separating creative and claims-review expenses from data-processing expenses and asking which stage changed outcomes or continuity.

10. Separate the fine from lasting business conditions

The fine’s basis was Xandr’s turnover, not Microsoft’s during the relevant period.^[5]

For business valuation, separating a one-time payment from later revenue conditions is useful. Illustratively, a company could incur a one-time cost, improve transaction conditions and preserve recurring revenue; the payment alone would not determine the change in value. Conversely, a small one-time cost could accompany a lasting deterioration in revenue or expenses. Analysis that ends with the fine leaves unresolved what the business can do after responding.

Accounting recognition, payment timing and future operating costs also need to be separated. Cash payments do not necessarily fall in the same period as the expense; actual treatment depends on company disclosures and applicable standards. Looking only at profit excluding a one-time item can omit its cash effect. Repeating that item as though it were an annual normal expense can understate recurring earnings. The denominator for a valuation ratio should be selected after tracing revenue, profit and cash.

Quantifying an effect on share prices or business value requires the relevant operation’s scale, margin, transaction changes, investment in the response and current procedural position. The announcement alone does not compress those factors into one number. Converting a fine at a current exchange rate is different from valuing future earnings. The same fine has different significance when comparison companies or periods differ. Before ranking multiples, align the business and earnings being valued.

11. The lag between old processing and post-acquisition assessment

The fourth analysis concerns the gap between processing, a change in ownership and publication of a decision. An acquirer gains future technology and customer relationships, while also considering problems from earlier operations that could emerge later. Material describing current products may not explain which information was sent to which recipients in the past. Assessing future commercial potential separately from the ability to explain past processing therefore has value.

Pre-transaction diligence costs should not be compared only with the size of the business acquired. Consider a company whose old transactions require further explanation. Uncertainty about staff, external expenses and access to material can make the timing of that burden hard to assess. The buyer compares spending on diligence with leaving uncertainty about later work unresolved. Unlimited expense cannot remove every question, so the scope of investigation needs to be linked with the acquisition terms it could change.

Allocation of liabilities, warranties and indemnities requires the individual transaction’s final agreement. A chronology alone establishes neither that pre-acquisition conduct cannot affect the buyer nor that the parent pays the entire fine. Connections between the legal entity, owner and payment obligation depend on law, contract and procedure. A change in ownership should not be expanded into a claim that all questions of responsibility have been settled.

The lag also matters in market analysis based on publication dates. The date of new information is different from the date of the underlying transactions. Explaining old share prices or advertising revenue with a current decision can introduce information participants did not know at the time. A point-in-time approach helps prevent this confusion. Old conduct can produce meaningful news; the task is to identify the information newly added to current assessment separately from the historical operations.

Aligning publication time and historical information — Free guide and educational time-alignment tool for avoiding hindsight.

12. Counterarguments and evidence that would change the analysis

A counterargument to a broadly bearish view is that different services support different demand. A market with dispersed buyers differs from one where a few buyers seek specific inventory when connection scope changes. Publishers with non-advertising income also have different capacity to absorb costs from those relying primarily on advertising. Examining demand composition and alternative income explains business differences better than applying one percentage change from an industry label to every company’s earnings.

The proposition that added costs inevitably raise prices also has counterarguments. Shared infrastructure or simpler processing might reduce costs enough to contain the total burden, even if checking work increases. Limiting information to what an advertising objective needs might simplify operations. Strong competition could prevent price increases and force a provider to absorb costs in its margin. Costs, prices and margins consequently require separate observation.

The proposition that larger businesses inevitably win assumes that reusable fixed costs dominate. If checking requirements differ by publisher or purpose, individual responses may grow with volume and weaken scale benefits. A smaller service with a narrower scope might compete by offering simpler transactions suited to particular demand. Equating size with capability overlooks that competitive opening. Reducing transaction complexity can itself be a business strength.

Evidence that would change the analysis is specific: procedural decisions changing responsibility, changes in actual processing scope, operations that match processing with consent, transaction volume and prices, and cost classifications. Stable revenue despite fewer recipients would weaken the hypothesis of dependence on broad connections. Individual-response costs growing with volume would require revising the fixed-cost-spreading model. Each scenario is more useful when paired with evidence that would support it and evidence that would overturn it.

13. The next procedural steps and places to look in disclosures

The appeal period is three weeks from receipt. The receipt date, a final court outcome and payment have not been confirmed.^[2]

Tracking procedure requires formal information on whether an appeal has been made, what the decision covers and its effect. The business implications differ if an outcome changes role assessment, consent assessment or only the fine. A change in the sum does not necessarily restore all earlier processing conditions. Conversely, a change in processing assessment can reveal new information about the company’s role that a focus on money alone would miss.

In company disclosures, separate a one-time sanction expense from operating changes. Information about recipients, data types and the correspondence between consent and processing helps examine revenue channels. Lower volume with higher prices and preserved profit differs from stable volume with greater checking expense. Beyond a short statement that a response is complete, evidence of which processing changed and how makes the economic assessment more concrete.

For advertisers and publishers, observation points include renewal fees, available features, usable information and migration costs. Without these, it can be difficult to tell whether the intermediary absorbed the burden or transferred it to counterparties. There is no need to invent dates for information whose publication has not been scheduled. Updating revenue and cost hypotheses when formal procedural information and actual transaction terms emerge is more useful than following dates alone.

14. Assess transaction reach together with the ability to explain it

Transaction reach and the ability to explain it are not measures that can be maximised separately and simply added together. Explaining broad transactions costs money; narrowing them to simplify operations can sacrifice demand. Comparing retained profit requires matching revenue and costs over the same period and identifying the burden accepted to win particular demand. Information volume and recipient lists alone do not reveal that trade-off. The purpose of this cost analysis is to incorporate operating conditions into assessments of growth.

Advertisers measure the price of outcomes. Publishers measure receipts after fees. Intermediaries measure the benefit of connecting participants against the cost of managing processing. Those interests overlap without being identical. A change in responsibility becomes market news by tracing who performs additional work and where the cost travels. Assuming that every company loses by the same percentage cannot explain demand shifts or differences between services.

After transaction conditions change, users and advertisers may take time to move elsewhere. Renewal timing, rebuilt connections and preparation of alternatives can make short-term profit and longer-term demand move differently. A first quarter alone may be too early to conclude that all added costs were absorbed or that demand was lost. Aligning transaction conditions with costs and then following subsequent demand shifts offers a way to assess durability.

This is reporting based on public material and conditional business analysis, not legal advice for an individual case or a recommendation to trade a financial product. Assessments of legality, contracts and investment value require the relevant processing and conditions to be examined individually. The hypothetical figures illustrate economic relationships; they are not actual company results, advertising prices or forecasts of market size.