5 Identity Resolution Case Studies: Real Results From Real Campaigns

Identity resolution produces dramatically different outcomes depending on how it is implemented. The technology itself — matching anonymous visitors to known individuals — is the table stake. What separates a program that returns 3 to 10 times on ad spend from one that produces a list of semi-relevant company names is how the identified data is layered, filtered, and activated.

The five case studies below come from Cursive campaigns across different industries and use cases. Each one shows a specific problem, the specific configuration that solved it, and the specific outcome. The through-line across all five: identity resolution alone is not the answer. Identity resolution combined with URL-stage scoring, noise suppression, and accurate buyer-stage classification is what moves the needle.

Case Study 1: B2B SaaS — Audience Accuracy from 12% to 94%

The problem: A B2B SaaS company running an account-based marketing program was sourcing intent data from a legacy co-op intent platform. The signals were keyword-based — aggregated from third-party publisher networks showing which companies were consuming content around their category terms. When sales reached out to the top-scored accounts, 12% showed genuine purchase behavior. The other 88% were noise — accounts that had appeared in the intent co-op data for reasons unrelated to an active evaluation.

The solution: Cursive deployed three layers in combination. First, the R3 industry cohort model was trained on the client's specific ICP — a SaaS-specific cohort that distinguished between companies researching the client's category and companies that merely showed up in broadly-defined keyword matches. Second, R4 noise suppression was applied, which removed 202,000 individual visit signals that did not meet the cohort's behavioral criteria. Third, R2 URL-level buyer-stage scoring was layered on top of the remaining signals to classify accounts by their specific stage in the evaluation process.

The result: Audience accuracy rose from 12% to 94%. Nearly every account the sales team contacted was in active evaluation — not in passive category awareness, not in competitive research for an unrelated project, and not a noise signal from the co-op data.

The URL-level breakdown tells the fuller story. Before the Cursive deployment, 8% of intent signals were at comparison or pricing pages — the pages that indicate an account is building a vendor shortlist. After R3 and R4 filtering, 73% of the remaining signals were at those same pages. The suppression did not just reduce volume. It concentrated the remaining signal into the accounts most likely to convert.

The underlying mechanism here is important for any team evaluating intent data quality. Co-op keyword data measures category consumption, not purchase intent. An account reading ten articles about a software category over six months scores high on keyword co-op data but may have zero purchase intent — they might be a journalist, a researcher, a student, or a company in a completely different vertical that uses the same terminology. The R3 cohort model learns to distinguish these patterns. R4 suppression removes the noise at scale. The result is a smaller but dramatically more accurate audience.

Case Study 2: Automotive Dealership — 22% Audience Reclassified as Active Buyers

The problem: A regional automotive dealership group was receiving intent signals on visitors "interested in automotive." So was every other automotive business in their market. Generic category interest is the lowest-value form of intent data — it tells you nothing that you could not infer from the fact that someone visited an automotive website. The dealership needed to distinguish browsers from buyers, and browsers in general from buyers in the next 30 to 90 days.

The solution: Cursive deployed URL-stage scoring calibrated to the dealership's specific site architecture. Individual pages were classified by purchase stage: vehicle landing pages (awareness), vehicle comparison pages (consideration), financing calculator pages (evaluation), and trade-in value pages (high intent / near-term purchase). Visits to evaluation and high-intent pages triggered immediate WaveState classification. Visitors were then identified by name where possible and routed to the sales floor.

The result: 22% of the dealership's site audience was reclassified as active buyers — visitors who had reached evaluation or high-intent pages during their session. This cohort was recommended for immediate outreach within 48 hours of identification, consistent with the WaveState trigger's optimal outreach window.

The significance of this result is not just the 22% figure. It is the precision it enables. Before URL-stage scoring, every visitor to the dealership's site was treated as roughly equivalent — someone who showed up got a generic follow-up ad. After URL-stage scoring, 22% of visitors could be identified as near-term purchase candidates and routed to direct, personalized outreach while the remaining 78% received appropriately lighter-touch nurture. The same total visitor pool produced radically different activation strategies for different segments based on page-level behavior.

Case Study 3: EdTech — 5.8x Audience Growth via Competitor Displacement

The problem: An online learning platform targeting enterprise L&D (learning and development) buyers faced a significant market disadvantage: a primary competitor had a three-year head start in audience development and brand recognition. Building an owned audience organically from scratch would take years. The client needed a faster path to a competitive audience that was already in-market.

The solution: Cursive deployed competitor displacement signals — identity resolution combined with behavioral signals indicating that a visitor was actively researching the client's primary competitors. This included visitors arriving via branded competitor search terms, visitors who had recently engaged with competitor content in the intent network, and visitors who showed comparison-page behavior suggesting they were building a vendor shortlist that included the competitor.

These competitor displacement signals surfaced 2,400 accounts actively evaluating alternatives to the incumbent platform. These accounts were flagged for specific campaign sequencing: a competitive differentiation message designed for buyers who were already familiar with the category and already in evaluation mode, not a top-of-funnel awareness message for someone who had never heard of the space.

The result: 5.8 times audience growth over six months, driven primarily by the competitor displacement cohort converting into engaged prospects. The key insight here is that audience building does not have to start from zero if you can identify accounts that are already evaluating your category and specifically looking at whether there is a better option than the incumbent. Those accounts represent the highest-velocity segment in any market — they are already sold on the category, already have budget considerations underway, and are actively looking for a reason to switch.

The right message for this cohort is never the same as your standard acquisition message. It needs to directly address why you are better than the specific alternative they are evaluating — not in a generic "we're better because" format, but with specific points of differentiation calibrated to the known complaints about the competitor and the specific outcomes the buyer cares about.

Case Study 4: Event Management SaaS — 34% Ad Spend Efficiency Gain via Suppression

The problem: An event management SaaS company was running paid acquisition campaigns across Google and LinkedIn. Performance was declining. Upon auditing the campaign audiences, they identified a structural problem: a significant portion of their ad impressions were being served to their own existing customers — users who were logged into the platform, clicking ads, and inflating engagement metrics without any possibility of converting to a new customer. Budget was being consumed by remarketing to people who had already converted.

The solution: Cursive's identity resolution was applied to the client's site traffic to build a real-time suppression list. As existing users visited the site — logged in or not, from any device — they were identified and flagged for suppression in active advertising audiences. The suppression list was updated continuously and synced to paid campaign audiences weekly.

The result: In week one of the suppression deployment, 187,000 existing-user visits were identified and removed from the active ad audience. Ad spend efficiency increased by 34% — the same budget was now reaching only net-new prospects rather than a mix of prospects and existing users.

The suppression use case is often the fastest path to demonstrable ROI from identity resolution because it addresses an active, measurable problem — wasted ad spend on audiences that cannot convert — rather than speculative opportunity. A team that deploys suppression first can fund the rest of their identity resolution program from the recovered efficiency within weeks.

The event management client found an additional benefit: their conversion rate metrics became significantly more accurate once existing users were removed from the audience. Conversion metrics that had appeared inflated by user re-engagement corrected to reflect true acquisition performance, which allowed the marketing team to make better channel allocation decisions going forward.

Case Study 5: Mass Tort Law Firm — 3x Outreach Conversion via Qualifying Behavior Filtering

The problem: A mass tort plaintiff recruitment firm was using broad keyword intent data to identify potential claimants for active litigation campaigns. The category keywords — product names, medical terms, symptoms — returned enormous volumes of signals, but the overwhelming majority were researchers, journalists, concerned family members, and people with tangential connections to the topic rather than actual potential plaintiffs. Outreach to the full keyword-matched audience produced low conversion and high cost per enrolled plaintiff.

The solution: Cursive deployed URL-level qualifying behavior filtering calibrated to the firm's specific case criteria. Pages were classified by the specificity of the qualifying behavior they represented: general educational content about the product or condition (low qualifier), symptom-matching pages (medium qualifier), diagnosis and treatment history pages (high qualifier), and plaintiff intake or legal rights pages (maximum qualifier). Only visitors reaching high and maximum qualifier pages were surfaced to the outreach team.

The result: 31% of the identified audience was classified as active researchers exhibiting qualifying behavior at the high or maximum level. Outreach to this filtered cohort produced a conversion rate three times higher than outreach to the unfiltered keyword-matched audience — because every person the team contacted had already demonstrated the specific research behavior that predicted plaintiff qualification, not general curiosity about the topic.

The mass tort case study illustrates a principle that generalizes across many industries: the most valuable signals are not the most common ones. A keyword match is common. A visit to a symptom-matching page is less common but more specific. A visit to a plaintiff intake page is rare but highly predictive of the exact outcome the firm is recruiting for. URL-level filtering concentrates outreach on the narrow band of behavior that actually predicts the desired outcome rather than the broad band that correlates loosely with the category.

Four Themes Across All Five Case Studies

These five campaigns are in different industries with different goals. But four principles appear in each one.

1. Identity Resolution Returns 3 to 10 Times on Ad Spend

The B2B SaaS accuracy improvement, the automotive buyer reclassification, the competitor displacement audience growth, the suppression efficiency gain, and the mass tort conversion improvement all represent significant positive ROI relative to the cost of the identity resolution deployment. The range is wide because use case and activation strategy matter enormously — suppression produces faster and more measurable returns than targeting because it addresses active waste rather than opportunity.

2. URL-Level Scoring Matters More Than Keyword Scoring

Every case study here benefited from URL-level page classification over simple keyword intent matching. Keywords tell you what topic someone is interested in. URLs tell you what stage of the evaluation process they are in. Stage information is more predictive of purchase behavior than topic information because the same topic can be researched by people at every stage of awareness — from curious first-time readers to buyers in final contract negotiation.

3. Suppression Is as Important as Targeting

The event management case study is the most direct illustration, but suppression logic appears in the B2B SaaS case too — the 202,000 suppressed signals were the mechanism that raised audience accuracy from 12% to 94%. Knowing who not to target is frequently as valuable as knowing who to target. Any identity resolution program that focuses only on the targeting side and ignores suppression is leaving significant efficiency on the table.

4. Feedback Loops Compound Over Time

Each of these programs improves over time as the identity resolution system accumulates more behavioral data about what actual converters look like on your specific site. The R3 cohort model trains on your traffic. The suppression list grows. URL classifications are refined based on which page types actually predict conversion versus which ones attract browsers. The first month of an identity resolution program will underperform the sixth month — not because the technology changed, but because the model learned. This compounding dynamic is why the attribution moat widens over time for teams that start early.

What These Case Studies Do Not Tell You

Every case study shows the outcome. What is harder to show is the work that precedes the outcome. The B2B SaaS accuracy improvement required careful ICP definition so that the R3 cohort model had the right training criteria. The automotive buyer reclassification required a thoughtful page taxonomy — someone had to define which pages indicated which stage. The competitor displacement campaign required messaging that was genuinely better than the generic outreach the competitor's churning customers typically received.

Identity resolution is not a magic layer you apply on top of existing programs to improve results automatically. It is a data layer that allows better decisions — about who to target, who to suppress, what to say, and when to say it. The quality of those decisions determines the quality of the outcome.

The teams that extract the most from identity resolution treat identified data as the beginning of a workflow, not the end of one. The identified visitor's name and email are inputs to a campaign that has been thoughtfully designed around what that person is researching, what stage they are in, and what they need to hear next. That design work is what turns a 70% identification rate into a 94% audience accuracy rate, a 3x conversion improvement, or a 34% efficiency gain.

Cursive's free traffic estimate shows how many named leads your site is currently losing before you commit to anything. The Super Pixel installs in under 10 minutes and begins building your identification data immediately. The results above represent the range of outcomes available once that data is layered, filtered, and activated with the right configuration for your specific use case.