With security tightening on personal data we are seeing more and more helpful ways to control the data flow from our devices to platforms and companies having the full picture about our lives and movements.
Looking at the settings on my iPhone I could see that Safari’s privacy report was supporting the efforts of specific sites which it is repressing/ blocking from gaining data and activity from my phone. A bit like a double or triple layer of security for my own privacy.
As consumers, we often accept cookie banners, app permissions and privacy notifications without giving them much thought. Yet, behind those settings is a bigger question: what happens to business performance measurement when people have more control over their personal data?
For businesses that rely on website analytics, digital advertising and conversion tracking, this is becoming an increasingly important consideration.
If a browser prevents a third-party tracker from following a user, or a visitor declines analytics cookies, some of the information traditionally used to measure marketing performance may no longer be available.
But does less visible data mean poorer performance, or do we need to simply approach the data in a different way or look for clues in the pattern of the data to read the performance?
To understand the blockers of the content, we need to understand what the devices are actually doing, and what our devices are actually protecting.
Privacy settings give us greater control over how our personal information is accessed, collected and shared.
As outlined in the Unanswered guide to accessing and managing privacy settings, devices and platforms offer controls for managing access to sensitive information, including location, contacts, cameras and microphones. Users can also review privacy settings within their browsers and online accounts.
On an iPhone, there are several relevant controls:
Safari Privacy Report: Shows cross-site trackers that Safari has identified and prevented from profiling browsing activity.
Prevent Cross-Site Tracking: Helps restrict the ability of third parties to track activity across different websites.
App Tracking Transparency: Allows users to decide whether apps can track their activity across other companies’ apps and websites for advertising and measurement purposes.
Location and app permissions: Allow users to decide which apps can access particular types of personal information.
Apple explains these protections in its Safari privacy overview and App Tracking Transparency guidance.
The important distinction is that these controls do not necessarily prevent a website from measuring activity on its own pages. Instead, they can restrict particular forms of tracking, data sharing and cross-site identification.
For consumers, this is a positive development. For businesses, it changes the conditions under which measurement takes place.
Why privacy settings matter for business performance
Digital performance measurement has traditionally relied on connecting activity across different touchpoints.
A business might want to understand whether someone saw an advert, clicked through to its website, returned a few days later and eventually completed a purchase. Analytics and advertising platforms use different combinations of cookies, identifiers, event data and other signals to help connect these interactions.
However, when tracking is restricted, or consent is not granted, some of those connections become harder or impossible to establish. This can affect several important performance metrics.
Website traffic: Some visits may still be recorded, while other activity may be unavailable to particular analytics tools because of consent choices, browser restrictions or implementation limitations.
Conversion rates: If conversion events are not consistently observed, the reported number of conversions may differ from the actual number of completed purchases or enquiries.
Marketing attribution: A business may know that a customer completed a purchase without being able to confidently identify which advertising interaction contributed to it.
Customer journeys: It can become harder to connect multiple visits or devices to the same person, making journeys appear shorter or more fragmented than they really are.
Audience measurement: Advertising platforms may have less information available to build, analyse or measure audiences across different websites and apps.
The result is a potential gap between what happened and what a performance tool can observe.
That distinction is crucial: a measurement gap is not automatically a performance problem.
What should we expect from performance measurement tools?
As privacy expectations evolve, businesses need to understand what their tools can measure reliably, what they can estimate and what remains unknown.
Should we be aware of the considerations, and how can we spot these? I have noted down a few below.
Expect gaps in observed data
No single analytics platform can guarantee a complete view of every customer interaction across every browser, device and marketing channel.
Some activity will be directly observed. Other activity may not be available because of privacy choices, technical restrictions or missing data.
Performance teams should understand these limitations and avoid treating every dashboard figure as a perfect representation of reality.
Expect more modelling and estimation
Some platforms use modelling to help account for missing information.
For example, Google Analytics 4 offers behavioural modelling in eligible Consent Mode implementations. It can estimate aspects of behaviour for users who decline analytics cookies, using patterns observed among similar users who consent.
Google explains the approach in its documentation on behavioural modelling for Consent Mode.
This can help businesses understand broader trends when some data is unavailable.
However, modelled figures are estimates, not directly observed interactions. Their availability and accuracy depend on eligibility requirements, implementation quality and sufficient data.
Businesses should therefore understand which reports contain observed data and which include modelled estimates.
Expect attribution to become less precise
Marketing attribution attempts to explain how different interactions contribute to a conversion.
When tracking signals are missing, assigning credit to a particular campaign, channel or advert can become more difficult.
A last-click report, for example, may show the final recorded interaction before a purchase without capturing the full journey that led to it.
This means businesses should be cautious about making major budget decisions based on a single attribution model.
Where appropriate, consider combining analytics with CRM outcomes, finance data, customer feedback and controlled experiments.
Expect differences between platforms
It is common for website analytics, advertising platforms and internal business systems to report different figures.
They may use different attribution windows, reporting definitions, consent signals, time zones and methods of counting conversions.
Privacy restrictions can add another source of variation.
Rather than assuming one platform must be correct in every respect, businesses should understand what each tool measures and establish a consistent source of truth for important business outcomes.
How can businesses adapt without compromising privacy?
The answer is not to find ways around people’s privacy choices. It is to build a measurement approach that works within them.
Review your tracking and consent setup
Check that your cookie consent platform, analytics tags and advertising tags behave according to the visitor’s choices and applicable requirements.
For businesses using Google products, review whether Consent Mode is correctly configured and whether the appropriate consent signals are being communicated.
Measure outcomes that matter
Focus on meaningful business results rather than relying exclusively on traffic and engagement metrics.
Depending on the business, these might include completed purchases, qualified leads, revenue, customer retention or cost per acquisition.
Where possible, reconcile analytics reporting against reliable operational records, such as completed orders in an ecommerce system or qualified enquiries in a CRM.
Separate observed figures from estimates
Make it clear when a report contains measured events, modelled conversions or other estimates.
This helps stakeholders understand the level of confidence they should place in the figures and reduces the risk of presenting estimates as exact counts.
Use multiple measurement methods
Website analytics remain useful, but they should not be the only source of insight.
Depending on the business, useful complementary methods include:
CRM and sales data to confirm business outcomes.
Marketing experiments to test whether campaigns produce incremental results.
Customer surveys to understand how people discovered a business.
Consistent financial reporting to evaluate revenue and profitability.
Trend analysis to identify changes over time.
These methods will not eliminate every measurement gap, but they can provide a more rounded view of performance.
Review your privacy settings and policies regularly
Privacy settings, browser behaviour and analytics tools evolve. Businesses should periodically review their tracking implementations, documentation and reporting assumptions.
For individuals, the same principle applies: review which apps have access to your information and understand what your browser is doing to protect your privacy.
The goal is not to collect as much data as possible. It is to collect and use appropriate data responsibly, with a clear understanding of its limitations.
A more useful way to think about performance
Perhaps the biggest change is that businesses need to distinguish between three different questions:
What actually happened? Did a customer buy, submit an enquiry or complete another valuable action?
What could our tools observe? Which interactions and conversion events were recorded?
What can we reasonably infer? Where can modelling, experiments or other evidence help us understand the activity we could not directly observe?
A business may have strong sales despite a decline in recorded conversions. Equally, it may have excellent tracking and poor commercial results.
The quality of a measurement system depends not only on how much data it collects, but also on whether it provides a credible explanation of what is happening in the business.
Better privacy requires better measurement
Looking at my iPhone’s Safari Privacy Report reminded me how much control people now have over their digital information.
The websites and trackers listed there represent more than a technical browser feature. They illustrate a broader shift towards giving individuals greater control over how their activity is collected and used.
For businesses, that shift presents a challenge, but also an opportunity. We need to interpret analytics more thoughtfully, be more transparent about uncertainty, and be more deliberate about the outcomes we measure.
A decline in reported traffic or conversions should prompt investigation, not an automatic assumption that performance has deteriorated. Similarly, a dashboard showing positive results should not remove the need to validate those results against real business outcomes.
Ultimately, effective performance measurement is not about tracking every individual at all costs.
It is about making sound decisions using the information we can responsibly collect, the evidence we can verify and the estimates we understand.
The future of performance measurement will belong to businesses that can respect customer privacy while still making confident, evidence-based decisions.
If you would like to read more, here are some useful sources to read further:
Unanswered: How to Access and Manage Privacy Settings — the original guide you supplied.
Apple: Safari Privacy — explains Safari’s cross-site tracking and fingerprinting protections.
Apple Support: App Tracking Transparency — explains users’ choices around app tracking.
Google Analytics: Behavioural Modelling for Consent Mode — explains how missing analytics data may be modelled.