Marketing Measurement Moves That Hold Up When Tracking Gets Fuzzy
Tracking marketing performance has become increasingly difficult as privacy regulations and platform changes reduce data visibility. This article presents 27 practical measurement strategies recommended by industry experts to maintain accurate attribution and ROI analysis. These tactics help marketers build reliable tracking systems that work even when traditional cookie-based methods fall short.
Forecast Organic Results With Audits
I have adapted by shifting measurement toward first-party SEO metrics and an SEO forecasting tool to make decisions without relying on cross-device signals. I focus on organic traffic growth, keyword ranking improvements, and lead generation as the primary indicators of channel impact. One simple change that improved trust was using the forecasting tool together with a detailed SEO and content audit to translate ranking and traffic changes into projected leads. That single adjustment made channel contributions clearer during the sales and planning process.
Combine Self-Reported And Platform Attribution
Marketing attribution has, honestly, gotten more messier over the last few years. Between the privacy updates, cookie restrictions, and people moving across multiple devices before they convert , the old " last-click tells all " thinking just doesn't really work anymore.
The biggest shift we've made is moving away from obsessing over perfect attribution and focusing more on blended performance indicators. Instead of asking 'which single ad caused this sale?', we look at how channels work together to influence demand over time. SEO, paid media, email, PR, and social all play different roles in the customer journey, and trying to isolate them too aggressively often leads to bad decisions.
One change that helped us get way more trustworthy insight was improving first-party data collection. We started asking leads directly how they found us, using smarter enquiry forms and a few onboarding conversations, then we cross checked that with what the platforms reported rather than leaning only on analytics tools . It sounds simple, but mixing what humans say about attribution with platform tracking gave us a much clearer view of actual channel impact.
And kind of ironically, as tracking gets less precise, good marketing has become more about understanding customer behaviour and less about chasing dashboards.
Tag By Message And Intent
What improved our measurement most was not another tool but a change in discipline. We started tagging campaigns based on message angle and audience intent instead of just platform and placement. When tracking gets weaker channel names alone do not explain performance clearly. Understanding what promise brought people in often tells us more than the device or browser path they took.
That one change gave us a clearer view of impact across campaigns. We saw that some channels worked well only with a specific narrative while others brought low confidence traffic. This helped us compare campaigns based on quality instead of surface level metrics. When attribution is incomplete better classification at entry often works better than trying to rebuild every missing click later.
Direct Ads Straight To Dial
The single change that made my channel attribution honest again was tying every paid impression to a phone call instead of trying to reconstruct a click path that gets shredded by Safari, iOS, and ad blockers.
We run paid spend across Google PMAX, Meta, and email at our small Hill Country resort. Browser-cookie attribution fell apart two years ago. Every dashboard told a different story, and we kept rebudgeting based on the loudest one. So we put a tracking phone number on the website and on every paid creative, fed it through CallRail, and watched what people actually did when they were ready to book. Calls don't lie. They show up tagged with the source, the keyword, and the time on the page before the dial. Two months in, we found that one campaign we were about to kill was actually driving 40 percent of our weekend bookings via phone — the GA4 dashboard had it close to zero because users called instead of clicking through.
If you sell anything that has a real human conversation in the funnel, build phone attribution before you build another dashboard. It's the cheapest measurement upgrade most operators are sleeping on.
Segment Subscriber Lists Per Origin
I started segmenting our email lists by source and tracking conversions separately. Third-party data is getting so unreliable that I needed something concrete. Since we own our email data, it tells the real story of where customers come from. Now I can see exactly which channels drive sales instead of guessing. If you want accurate numbers, stop relying on messy tracking and start looking at your email lists by source. It makes a huge difference.
If you have any questions, feel free to reach out to my personal email
Enforce Shared Reporting Governance Monthly
The adjustment was less technical than operational, because measurement quality usually fails through process first. Cross-device loss exposed how loosely many teams define attribution, conversions, and qualified demand. Better insight came from tightening governance before buying another analytics layer. We standardized naming, ownership, and validation rules across campaigns, reports, and revenue stages.
One simple change created the biggest jump in trust. Every month, marketing and sales review a shared source-of-truth report with exception checks. Mismatched lead counts, duplicate opportunities, and impossible conversion rates are investigated before conclusions. That discipline reduced reporting fiction and made channel impact clearer than any attribution model promise.
Judge Acquisition By Cohort Revenue
Cross-device tracking in SaaS is a mess, so I stick to subscription cohorts by channel. I track where trials start and, more importantly, which sources actually drive paid months later. When we stopped obsessing over clicks and just looked at revenue by source, we found some channels were bringing in way more valuable users than we thought. Honestly, a basic spreadsheet tracking trial-to-paid rates usually tells you more than the expensive tools do.
If you have any questions, feel free to reach out to my personal email
Ask Customers Directly On Cards
Tracking how customers discovered us across different devices and platforms felt increasingly unreliable the more we tried to depend on digital analytics alone. Our dashboard numbers looked confident but they rarely matched the real conversations happening at market stalls, pop up exhibitions and community upcycling events where most of our genuine connections were actually being made. The most honest measurement shift we made was beautifully simple. Every product leaving our workshop carried a small handwritten card inviting customers to share in their own words how they first found Dwij. Not a dropdown form, not a digital survey, just one genuine open question on recycled paper. Within four months that simple human question revealed that 67% of our most loyal customers had discovered us through personal recommendations and community conversations that no digital tool had ever captured or credited. We stopped trusting dashboards completely and started treating direct human conversation as our most accurate and revealing measurement instrument.
Assign Channel-Specific Phone Numbers
I've been navigating the tracking challenges at A-S Meds for the past year, and honestly, it's pushed us to get more creative with how we measure our marketing efforts. When we noticed our analytics weren't painting the full picture anymore, especially for our healthcare provider campaigns, we knew something had to change.
The biggest shift we've made is moving away from relying solely on last-click attribution. We used to see a doctor or clinic manager click through from an email, then come back later through organic search to place an order for surgical supplies or durable medical equipment. The old system would give all the credit to organic, completely ignoring that our email campaign started the conversation.
One simple change that's given us much better insight is implementing unique phone numbers for each marketing channel. I know it sounds basic, but hear me out. Medical professionals often prefer calling to discuss bulk orders for things like wound care supplies or mobility equipment. By assigning specific phone numbers to our email campaigns, social media ads, and direct mail pieces, we can finally capture those phone conversions we'd been missing.
We also started using more first-party data collection through our website forms. When healthcare providers request quotes or download our product catalogs for items like diabetic supplies or respiratory equipment, we ask how they found us. Simple, direct, and surprisingly accurate.
The phone tracking alone showed us that our email campaigns were driving three times more conversions than we thought. We were ready to cut that channel's budget, which would have been a huge mistake.
I'm not saying this approach is perfect. Cross-device behavior is still tricky to capture, and we're continuing to refine our methods. But trusting actual conversations over flawed digital metrics has transformed how we allocate our marketing budget at A-S Meds. Sometimes the simplest solutions really do work best.
Analyze Sales Recordings For Impact
Instead of relying solely on analytics platforms, we now include call recordings and marketing data in our measurement strategy. For each sales call we record, a team member digs through the data to identify which phrases correlate with a specific campaign, landing page, or advertisement.
An important insight came from a healthcare client. While attribution reports showed Facebook campaigns performing at a more typical level, prospects often quoted a sentence from the short-form video ad during consultations. While analytics give little credit to social channels, the sales calls were a different story.
This revolutionized the way we think about measuring impact. Especially when it comes to cross-device engagement, the path a buyer takes is anything but linear. The most useful feedback came when probed about what prospects actually remembered, rather than relying on vain software metrics. After we stopped treating attribution tools as the highest authority and started challenging their campaign results against real conversations with people, marketing measurement became something you could rely on.
Treat CRM As Source Of Truth
With tracking and attribution being one of the biggest nightmares for marketers, the real difference when measuring your strategies comes by understanding the playing field, knowing how and why each platform claiming the attributed sale or engagement for themselves does not match what your CRM says.
Reading each channel separately helps as a guide and you can surely optimise based on platform insights, but your CRM as a whole is your real source of truth. Being able to see the bigger picture is the differentiating factor for a marketer, as the most common scenario is that your audience and customers have more than one single touch point with your brand before committing to spend their hard earned cash with you.
Knowing how not to obsess about organic traffic declining, as it's happening for pretty much every website in the world with the rise of AI citations, is an important insight to keep it mind as the traffic you're getting comes with stronger signals of engagement.
In short, the channel impact should be seen on the organisational perfomance as a whole, the bottom line. An integrated marketing strategy doesn't work in siloes so your insights need to come from a holistic understanding on how each channel interacts with each other.
Deploy UTM-Tagged QR Codes Offline
I've definitely felt the pain of declining tracking reliability at Free QR Code AI. Cookie deprecation, iOS privacy changes, and ad blockers have made our traditional attribution models pretty shaky. We used to rely heavily on last-click attribution, but that approach just doesn't cut it anymore when you can't follow users across their phone, tablet, and laptop.
The single biggest change we made was shifting to UTM-tagged QR codes for our offline-to-online campaigns. I know that sounds self-serving given what our company does, but hear me out. Instead of hoping cross-device tracking would magically work, we started creating distinct QR codes for different physical placements and partner channels. Each QR code carries specific UTM parameters, so when someone scans it, we know exactly which channel drove that visit, regardless of what device they're using or what browser settings they have.
This gave us something we'd been missing: deterministic rather than probabilistic attribution. When someone scans a QR code from a specific trade show banner or partner flyer, that's a confirmed touchpoint. We're not guessing based on fuzzy device graphs.
We also started using unique landing pages for major campaigns rather than sending everyone to our homepage. Combined with the QR approach, this meant we could see clear paths from scan to conversion without relying on third-party cookies.
The insight was immediate and sometimes surprising. Channels we thought were performing well based on old models turned out to be overvalued, while some offline partnerships we'd almost cut were actually driving serious revenue. It changed how we allocate our marketing budget completely.
We've accepted that perfect tracking isn't coming back. But direct response tools like QR codes give us solid, honest data we can actually trust.
Prioritize Trends Over Perfect Precision
Call it personal growth, but as tracking has become less reliable I've moved away from trying to force perfect attribution. It just doesn't seem like it can be done as easily as it used to be. Now I tend to look much more on broader performance patterns instead of relying only on platform conversion reporting. Looking at trends in traffic quality, revenue movement, and conversion behavior over time gave us a more trustworthy view of channel impact even though it was a bit more labor intensive. We also simplified reporting around a smaller set of shared business metrics instead of debating differences between platforms, since it is hard to tell which one in particular should act as the north start. That reduced a lot of confusion. For me, the biggest lesson was that modern measurement is more about interpreting signals than expecting exact precision and that means a bigger scoop of data is necessary.
Launch Account-Based Loyalty Program
Since we started an email loyalty program at Japantastic, we can finally see what users are doing across their phones and computers. Because they log in everywhere, we connect the dots on their activity instead of guessing. If you are tired of losing track of customers when they switch devices, you should probably just get them to sign up for an account. It beats relying on cookies.
If you have any questions, feel free to reach out to my personal email
Extend Conversion Windows For Clarity
As a performance marketing agency its imperative to track for business outcomes, one direction we ought to take is to get our website infrastructure on 'server side tracking' but we decide to cross that bridge when we get there. Here is why we opted out off server side tracking, cost, complexity and maintaining cloud infrastructure.
Moreover Googles Smart Bidding Algorithm is retaining conversions paths for 6 months for better results. Unless you are missing over 20% attribution data this practice is time consuming and makes an expensive complexity slightly more manageable.
One challenge we encountered in UAE, they do not permit Google to track phone calls from Ads, hence all the conversion from phone ads or calls are not attributed regardless of standard server side tracking (sGTM), this needs an Offline Call Conversion Import.
and at the same time for E-Com clients its a mandatory practice for real time remarketing, data security and GDPR, this keeps the CRM in sync with ads conversion and hyper personalized email gets relevant context.
Pro tip to enhance conversion recall is to expand the conversion window to the maximum limit of 90 days.
Adopt Media Mix Models
The industry has spent the last decade changing between the "magic" of digital tracking and the harsh reality of data fragmentation. The current landscape of less reliable tracking is a significant challenge, but it is also an opportunity to return to more robust marketing measurement.
Poor Tracking Isn't a New Issue
While the focus is on the recent "death of the cookie", poor tracking isn't actually a new problem. We have long known that digital signals are imperfect. They struggle with cross-device journeys, and they have always failed to capture the offline world of TV, OOH, or word-of-mouth. The "tracking" we've relied on for years has often been a mirage of precision.
Single-Channel Measurement still works
Digital tracking still "works" if you have one very narrow goal: measuring one channel against one immediate action, such as a direct sale from a search click. However, as soon as you introduce multiple channels, the system breaks down. The flaw in last-click attribution is that it disproportionately rewards channels closest to the purchase. It ignores the long-term work of brand building activity that created the demand.
A Simple Change: Embracing the incremental impact of marketing
If you are looking for one simple change to get more trustworthy insights: Stop treating platform-reported ROAS as truth and start using Marketing Mix Modelling (MMM) for cross-channel measurement. At Linea Analytics, we encourage brands to move away from fragmented tracking and toward a cross-channel view. MMM is uniquely positioned for today's world because it doesn't rely on user tracking. Instead, it uses aggregate data to isolate the incremental impact of every marketing channel.
Why MMM is the Gold Standard for CMOs
MMM allows for a "like-for-like" comparison across all channels, be that digital or traditional. It does this by controlling for external factors like seasonality, economic shifts, and competitor activity, that's the incremental impact. Two key benefits of MMM include:
1) Measuring Long-Term Impact: We can capture the memory effect, ensuring that brand-building activity is fairly valued alongside short-term tactics.
2) Setting Budgets with Confidence: By calculating diminishing returns, we can show CMOs exactly where the next $$ should be spent to maximise revenue.
This isn't to say MMM can answer all questions, but using MMM alongside other approaches provides Brands the right ingredients to better marketing measurement.
Check Topic Trust In Comments
Most creator partnerships fail at the vetting stage, not the execution stage.
The standard vetting checklist is follower count, engagement rate, niche relevance, and brand safety. All four are necessary. None of them tell you what actually matters, which is whether the creator's audience trusts them on the specific topic you are paying for.
A creator with 200,000 followers in the fitness space might have a deeply trusted audience for workout programming and almost zero trust on supplement recommendations. Same audience, same creator, completely different conversion outcome. The follower count looks the same to a brand manager. The actual partnership outcome will not be the same at all.
The vetting step that prevented the most mismatches for us is what I call the "specific topic comment audit." Before signing any creator partnership, we go to their last twenty pieces of content and read the comments on the three pieces closest to the topic the partnership will cover. Not their highest-performing content. The content nearest the actual product or service angle.
What we are looking for is the language pattern in the comments. If the comments are mostly emoji reactions and "love this," the audience is engaged but not trusting that creator's expertise on the topic. If the comments are detailed responses where people share their own experiences, ask specific follow-up questions, or push back with their own data, that is an audience that treats the creator as an authority on that specific subject.
We did this for a B2B SaaS client doing a partnership with a LinkedIn creator in the productivity space. The creator had strong overall engagement, but when we audited the comments on their three most product-adjacent posts, the discussion was almost entirely generic affirmations. No specific questions, no follow-up workflows, no real engagement with the operational substance. We flagged the mismatch.
The client paused, picked a smaller creator instead, and that smaller partnership produced three times the demo bookings the larger one was projected to deliver. The smaller creator's comments on adjacent topics were full of people asking detailed setup questions, sharing their own stacks, and disagreeing constructively. That audience was already in the consideration mindset the client's product needed.
Measure Repeat Purchases From Database
We stopped tracking website sessions and started looking at repeat purchases from our car mat customers instead. The data in our CRM showed which channels brought buyers back and which only brought one-time customers. It completely changed how we spent our marketing money. If you're tired of unreliable browser data, just look at what your existing customers are actually doing. That's the data you can trust.
If you have any questions, feel free to reach out to my personal email
Track Page-Level Return Visits
At Pikadil, one simple change that helped was focusing more on repeat visits tied to specific pages instead of only last-click conversions. We noticed a lot of users discover a deal on one device, then come back later through another channel to actually complete the purchase. Looking at engagement patterns over time gave us a clearer picture of which channels were genuinely influencing buying decisions.
Rely On Fresh Review Volume
I stopped looking at stale reviews because they skewed my results. Now I wait for at least 50 reviews, with 80 percent from the last three months. This gives me a real sense of which channels people are using. It is a solid alternative to cross-device tracking, which is getting harder to trust these days.
If you have any questions, feel free to reach out to my personal email
Run Controlled Reproducible Comparisons
The practice of tracking hasn't necessarily changed course when it comes to differences between devices and browsers. The only thing it does now is stop making a show of being objective. At SearchTides since 2017, over 20 years in search marketing and LLMs, especially in the finance and healthcare sectors, perfection in attribution was always out of reach. It was just a prettier mess in the dashboard than elsewhere.
The last click, cookies, and session stitching - these things were only ever showing a partial view before. They have merely gotten more transparent about it. The demand for certainty from numbers that were never entirely sure remains constant, and is often required even in regulated spaces where funnel completion is impossible.
The paradigm has thus moved from attribution to that of signal directionality. This implies considering the data as an indicator of strength rather than proof of ownership. SEO impact is now measured through longer periods, just as LLM visibility is done. You have been aggregating weak signals rather than searching for one single source of truth. This approach is actually more palatable to financial clients than many think due to the way they operate.
One thing that worked much better than anything else was using controlled comparisons between different groups. Groups that had similar pages, similar intent clusters, and were compared for only one factor at a time. Comparisons between treated and non-treated groups over equal time periods. It filters out lots of noise across devices without improving tracking.
Single dashboards began to lose their credibility because they keep contradicting each other on a regular basis. Analytics, customer relationship management (CRM), search engine console - all showing different results. The shift toward privacy in the browser further made this discrepancy evident. SEO-driven traffic gets underestimated or attributed improperly, which leads to disputes based on reports rather than understanding.
This means that it becomes more important at this stage to be reproducible, rather than accurate. If something proves to be true after multiple repetitions, then the numbers don't really matter that much. The trend is far more important than any snapshot of accuracy.
Import Offline Outcomes With Click IDs
Cross-device and browser tracking keeps degrading, so we stopped relying on it as the source of truth. The single change that made the biggest difference was moving measurement server-side: every inquiry carries its Google Click ID into the CRM, and once sales qualifies a lead, that status is uploaded back to Google Ads as an offline conversion tied to the original click.
This survives whatever the browser does. Cookies dropping, ITP, iOS restrictions, none of it matters because the click ID is captured at form submission and lives in our database, not in a third-party cookie. Smart Bidding then trains on qualified outcomes from our CRM rather than on-page events that decay over time.
One technical note most Google Ads setups miss: Google uses three click identifiers. gclid is the default, gbraid is used for IOS, wbraid covers in-app. If your offline conversion action is set to count = "1" (the default), gbraid and wbraid uploads will error. You will need a separate offline conversion action with its count set to "Every" for it to accept gbraid and wbraid.
With one of our clients, fixing this signal alone produced a 3x increase in qualified leads over 2 quarters, without any other major campaign changes.
Purge Orphaned Tags Each Quarter
The first move we make is the opposite of adding tracking. We audit what is already running and delete what is not working. The measurement gets more trustworthy because the noise gets quieter.
The pattern we see on nearly every new client. A GTM container running Google Analytics, Hotjar, Microsoft Clarity, and Crazy Egg simultaneously. Three of those overlap on the same heatmap and session-recording job. Two of them are firing pixels left over from retargeting campaigns that ended in 2024.
One healthcare client had retargeting pixels firing for a paid campaign that wrapped two years before we audited the container. The pixels were still collecting page-level data and sending it to platforms the client had no BAA in place with. A page-speed hit and a privacy exposure that also made the measurement signal untrustworthy.
The simple change is a quarterly tag audit. Open the container, document every tag, document what it is supposed to do, document the last time it actually fired against the intended campaign. Delete the ones with no current owner.
Channel impact gets clearer because the channels you are still running stop competing with ghosts. Tools you do not use cannot mismeasure you. The unlock is subtraction.
Most teams think the measurement problem is data scarcity. The actual problem is unowned data exhaust nobody has audited in two years.
Trevor Gage, Director of Marketing, Webserv. Webserv works exclusively with behavioral health and addiction treatment centers on SEO, content, and earned/owned media.
Switch To Cookieless Plausible Analytics
Tracking has gotten less reliable for three reasons stacked together: ad-blockers, browser-level cookie restrictions (Safari ITP, Chrome's third-party cookie deprecation), and consent-banner opt-outs in the EU. On a small-business marketing audience like ours, the combined data loss can hit 30 percent or more in GA4.
The simple change that gave us cleaner insight: switching from GA4 to Plausible Analytics as the primary measurement tool. Plausible is server-side and cookieless, which means it captures every visit regardless of ad-blocker status, doesn't trigger consent banner requirements in EU, and doesn't sample data on the free tier the way GA4 does.
The trade is real. Plausible doesn't do user-level journeys, multi-touch attribution, or audience segmentation the way GA4 does. For us, that trade-off was worth it because we'd rather have honest pageview-and-source data than fancy reports built on incomplete data.
The other change that helped: stopped trying to do precise multi-touch attribution. Instead, we ask three simpler questions per channel: did the channel send measurable traffic this month? Did the traffic from this channel convert at our goal rate? Is the conversion volume from this channel large enough to justify the time it took? If yes to all three, the channel is working. If no to any, fix or drop.
The deeper lesson: trustworthy measurement at our scale comes from fewer, simpler metrics that survive the data-loss problem, not more sophisticated metrics that depend on tracking we can't reliably do anymore.
Manu Hampton,
Founder, The Stack Reviewer (thestackreviewer.com)
Monitor Affiliate Revenue By Source
I stopped obsessing over individual clicks on Way2Earning and started checking affiliate revenue by channel every week. Grouping traffic sources and watching the trends, especially during campaigns, actually works. Even with messy device tracking, I can see what brings in the money. If your data feels off, stop counting visitors and start looking at the results by source. You will figure out what is working much faster that way.
If you have any questions, feel free to reach out to my personal email
Make Branded Search Lift Your KPI
The shift that gave us the most trustworthy signal back was switching the primary KPI from last-click attribution to branded-search lift in Google Search Console.
We operate a B2C SaaS in the streamer-audience-growth vertical where iOS 14.5 ATT, third-party cookie deprecation, and now AI-search referrer opacity collectively broke our last-click model around mid-2025. Channels that GA4 reported as 60-day high-performers were showing zero correlation with revenue when we looked at cohort data manually. Channels we had nearly cut were quietly driving direct-traffic spikes 7-21 days after exposure.
The simple change: track weekly branded-impression count in Google Search Console alongside direct-traffic-to-homepage, segmented by week, and treat the lift in those two as the primary indicator of channel impact. Verify with a holdout-region test — pause the channel in one geo for 14 days, watch the branded-search and direct-traffic curves diverge from the live geo.
Why this works when click-path tracking does not: branded-search and homepage-direct are events where the user explicitly types or pastes your brand name. They cannot be cookie-blocked, browser-stripped, or ITP-degraded. They are the lowest-funnel intent signal that survives every privacy framework, and they correlate with revenue 6-10 weeks downstream at roughly 0.7 R2 in our data.
The follow-on operational change: switched paid-channel budget review from monthly attribution dashboards to a fortnightly branded-impression delta review. Channels that produced visible delta got renewed. Channels that did not stopped getting renewed regardless of what last-click reported.
Concrete result over the eight-month transition: reported channel ROI numbers dropped 30-40% on the dashboard (we had been overcrediting paid channels) but actual revenue stayed flat then grew, because budget started flowing toward channels producing real demand rather than ones intercepting demand that would have shown up anyway.
The rule that emerged: if your KPI can be measured 90 days after the user first encountered the brand, it survives the tracking collapse. If it depends on attributing a click to a session to a conversion within the same week, it is built on a measurement stack that no longer exists.
Triangulate Results With Consistent Codes
With tracking becoming less reliable across devices and browsers, I've had to shift away from depending on a single "source of truth" like last-click attribution or platform-reported conversions. Instead, I focus more on triangulating performance using multiple signals rather than one dataset.
The biggest change I made was simplifying how I measure success: I started tracking fewer metrics but tying them directly to business outcomes, especially lead quality and conversion actions, not just traffic or clicks. This means looking at how many qualified inquiries, bookings, or sales actually come from each channel, even if the attribution is not perfectly precise.
One simple but powerful adjustment was implementing consistent UTM tagging across all campaigns combined with a centralized dashboard that compares platform data (like Meta or Google Ads reports) against backend conversion data from CRM or booking systems. This helped highlight discrepancies and reduced overconfidence in platform-reported numbers.
The result was a more realistic view of channel performance. Instead of chasing "perfect attribution," I now focus on directional accuracy ;understanding which channels consistently bring high-intent users, even if the exact journey isn't fully traceable.
