Rideshare on Meta Ads Case Study: LegalCalls Approach

The Legal Calls Approach to Rideshare cases from Meta ads

Step 1: Meta Ads

Every case in our pipeline starts the same way: someone scrolling Facebook or Instagram. Our creatives and copy are built to reach women who were harmed during an Uber or Lyft ride and get them to raise their hand, all without tripping Meta’s policies. And because no two women react to an assault the same way, we don’t run one emotional note. Some creatives are informative and plain, and others lean into the feeling, so whether she’s angry, frightened, or still processing what happened, there’s a version that speaks to her. We run dozens, kill the losers, and scale the winners.

The Trifecta: Every ad we run is intended to achieve three core objectives.The first objective is clearly establishing who the ad is speaking to; women who rode as passengers in an Uber or Lyft. The second – conveying why they should care. This can take on different forms and styles to resonate with the intended target. The last objective is to clearly explain what the ad is about: a sexual assault that happened during the ride. Plenty of women take Ubers; the ones we need are the ones who were assaulted during the ride, so that’s what the ad leads with.

We’re not trying to stop everyone: An ad that grabs the entire platform is a losing ad. All it does is drain the budget on people who will never have a case, and it teaches the Meta pixel to go find more just like them. The only scroll we want to interrupt is the woman who reads it and thinks, that happened to me. Everyone else can keep going. We’re not in the business of talking anyone into a claim; we’re looking for the small number of riders who were actually assaulted, and letting the rest pass right by.

We stay inside Meta’s lines: Meta is strict about two things that both land on us, so we build around both. First, it won’t let an ad imply it knows something private about the person reading it, so we never say “you were assaulted.” Second, it restricts explicit sexual language outright. What’s left is third-person framing: “Women harmed during an Uber or Lyft ride may be eligible for compensation.” It gets the same message across without breaking a single rule. The real qualifying gets pushed to the landing page and the form, where we’re actually allowed to ask.

The pixel data feedback loop: The pixel is only as smart as the data we feed it, so we’re careful about what we send back. When our call center updates a lead’s status, a webhook pushes that status to Meta, which teaches it to chase real claimants instead of people who just like filling out forms. We do this in two stages. The first signal, our qualified event, fires once a lead answers the key questions the right way, including whether she was a passenger or a driver. That tells Meta we’ve found someone who is female, a passenger, and reachable, which is all we need to have a real shot at signing her. The second signal, our further-qualified event, fires later, once she has provided her ride receipt and finishes an intake, which teaches the pixel the deeper pattern of a claimant who actually qualifies for the lawsuit. One thing we’re deliberate about: a wrong answer on the form never kills a lead on its own. People tap the wrong option all the time, so a human still calls every lead, whether their answers indicate they are qualified or disqualified. The form points Meta in the right direction; our agents decide who’s real.

Step 2: The Landing Page

Once she taps the ad, the landing page takes over. It’s built on mobile first, and it has one job: confirm she’s in the right place and get an honest first read on whether she might have a case. Every piece of it, from the headline down to the order of the questions and the thank-you screen, is tuned to reduce friction without dropping the qualification bar.

Ad-to-landing-page congruency: The headline on the page picks up the exact language from the ad she just tapped, the same reference to being harmed during an Uber or Lyft ride. When the page echoes the ad that earned the click, she knows in a half-second she’s in the right place. Fewer people bounce, and more convert.

Mobile-first design: Almost every woman who reaches this page came from Meta on her phone, so it’s built for mobile first: quick to load, easy to thumb through, with a click-to-call button for anyone who’d rather just talk to a person than fill anything out. We run it through BrowserStack to see how it renders across every browser and device combination, Safari included, so it looks right no matter what device she’s holding.

The progressive form: We never drop the whole form on her at once. It’s one question per screen, and the order is deliberate. The easy, no-stakes questions come first (name, email, phone), and the heavier ones come later. By the time she reaches anything sensitive, she’s already answered three or four simple things and built a little momentum, so she’s far more likely to finish than if we’d opened with the hard question. Every extra field costs conversions, so we only ask what actually moves her toward qualifying and nothing that doesn’t.

The category question is built for honest answers: We never ask her to confirm the worst word about the worst day of her life. A blunt “Were you sexually assaulted? Yes or No” does two things wrong at once: it’s leading, and it’s the kind of question a lot of women simply won’t answer on a form. So instead of a yes/no, we hand her a list to choose from: Verbal Harassment, Indecent Exposure, Inappropriate Touching, Physical Assault, Sexual Assault, Kidnapping. She isn’t declaring anything or typing the hardest sentence herself, she’s just selecting the option that matches what happened, which is far easier to do truthfully. Because the list runs from milder to severe, it never feels like an all-or-nothing trap, and there’s a free-text box if she wants to say more in her own words, though she’s never required to. Selecting a category is the gate, so there’s no dead-end “no” for anyone to hit. Every honest answer gets captured and contacted. The serious categories are the ones that really qualify and teach the pixel who to look for, while the lighter ones still get a call, because some of those turn into real cases once an agent actually talks to her.

We look like a law firm, not a lead mill: Given the subject, the page has to read instantly as a serious legal service. Everything about it, the branding, the language, the tone, signals a law firm helping survivors pursue a claim, not anything adult or exploitative. That keeps Meta’s classifier from misreading the sexual-assault subject matter and filing us under the wrong rulebook. And just as important, it earns the trust of a woman who has every reason to be careful about who she’s talking to.

The thank-you page sets up the call: Before she leaves the page, we tell her exactly what happens next so the call doesn’t catch her off guard. The screen lets her know our intake team needs to speak with her by phone to move things forward, that we’re calling right away, and that she should expect to hear from us within a minute. We tell her the caller ID will read “Justice for Women” so she recognizes the ring instead of sending it to voicemail, and we put a call-now button right there on the page. The most motivated women don’t wait for us at all. They tap it and call in themselves.

Step 3: The Call Center

The moment that form hits our system, the call center takes over. The goal is to give ourselves the best possible chance at making contact, and to treat every lead as if she’s already a qualified claimant until we learn otherwise. In this tort that second half matters more than usual, because part of the job is sorting the real cases from those that simply don’t qualify.

Speed to lead: The second a new lead comes in, it drops into our team’s Slack and an agent is calling within about 30 seconds. We want to reach her while she’s still expecting the phone to ring, not hours later when her interest has cooled and she’s had second thoughts about picking up. Every new lead gets called twice at once to signal this is real and it’s urgent. If she doesn’t answer, we leave a voicemail and follow it with a personalized text so she knows who reached out and why.

Personalized from the first touch: The form gives her a free-text box to describe what happened in her own words, and the women with real cases usually do. We use what she wrote to make the follow-up personal, so our text or voicemail speaks to her actual situation instead of reading like a generic script. It feels like it’s coming from someone who read what she said, because it is. Given what she’s writing about, that difference matters. It’s often the first sign to her that she’s reached people who are actually listening, not a call center working down a list.

Maximizing contact: Getting her on the phone is half persistence and half making sure our calls actually get picked up, so we work both. On persistence, our outbound cycle runs 21 days, with calls at varying frequency backed by a mix of text and email. The point is to catch her at a time and in a way she’s actually willing to respond to, instead of quitting after a couple of missed calls. On answer rates, we manage our caller ID two ways. Our numbers display as “Justice for Women” so she recognizes who’s reaching out, and we run a service that constantly monitors the reputation of those numbers and automatically remediates any that start to slip, so our calls keep landing as legitimate instead of getting flagged as spam.

Agent training: Training starts with sensitivity, since she may be recounting the worst thing that’s ever happened to her, and the way an agent handles that first conversation often decides whether she stays on the line. Agents also know the qualification criteria cold, so they can tell early whether a rider fits. Most of the training, though, is about fraud. Because a claim carries almost no risk for the person making it, the driver faces nothing and the claimant rarely sees a courtroom, this tort draws a steady stream of fabricated stories, and telling those apart from real survivors is the hardest part of the job. Agents learn the tells: accounts that shift or contradict themselves, stories that come across oddly thin or suspiciously over-rehearsed, and receipts that don’t hold up. On receipts, they’re trained on what Uber’s and Lyft’s legitimate formats look like (for example Uber One has a different receipt format than most Uber receipts), and on how to walk a claimant through the app and her email to produce the right one. When a red flag is noticed, the agent escalates the call to a manager.

Manager escalations: There are a handful of situations where a manager steps into the call, and they fall into a few buckets. Some are about helping the claimant through a sticking point, like when she can’t locate her ride receipt and needs someone to walk her through retrieving both the PDF and the email copy, or when she’s uncomfortable talking about what happened and a calmer, more experienced voice helps put her at ease. Others are about protecting the quality of the case. When an agent notices inconsistencies in a story, a manager takes over and works through those questions more directly to get to the bottom of it, and the same applies to any statute-of-limitations concerns, which a manager is better positioned to evaluate. And then there are the strong cases, where a claim looks like a good one we’ll be following up on. In those, a manager introduces themselves and shares a direct line, so the claimant knows her case is being taken seriously and has someone she can reach.

Notetaking and dispositioning: Every call is documented as it happens. Agents log what was discussed, where the claimant stands, and anything still outstanding, so each lead is worked to its fullest and nothing gets lost between calls. Part of this is keeping her status current, which is what feeds the signals back to Meta and tells us who to follow up with and when. A common example is a claimant who qualifies but doesn’t yet have her ride receipt, in which case we note it and follow up with her later to obtain it. Agents also flag anything worth a second look, whether that’s a possible red flag in her story or a question that hasn’t been fully answered, so it’s on the record for a manager or the QA team to pick up.

Step 4: QA

Once a case qualifies and the claimant signs, it moves to a dedicated QA team led by our fulfillment manager, Lorenia. Her job is to make sure every case we pass along actually holds up: that the ride was real, the story is consistent, the documentation is complete, and the claim fits the criteria. Given how much fraud this tort attracts, this QA step is important to the overall quality we deliver to our law firm clients.

Verifying the ride: The first thing QA does is confirm the ride itself was real. That starts with the receipt and the confirmation email from the app, which gets a fresh set of eyes to check them against what a legitimate one should look like (we have caught a few people submitting fake, AI generated receipts). From there, QA examines the receipt against her account of what happened, comparing the pickup and drop-off times and locations on the receipt to the story she gave at intake. When those details match, the ride holds up. If her story doesn’t fit the timing or the route on the receipt, that’s a sign the case needs a closer look before it goes any further.

Verifying the claim: With the ride confirmed, QA turns to the claim itself. First, they check that the type of assault she reported is one that qualifies under the criteria. Second, they confirm she reported the incident around the time it happened, whether that was to the police, a medical professional, a therapist, or someone close to her, or in some cases that she has a recording of it. Our agents collect this at intake, and QA confirms it’s documented and consistent with the rest of her file.

Completeness and background: QA also does a final pass to make sure the file is complete. That means all the documentation is present and in order. This includes her photo ID, the ride receipt, and the confirmation email. QA also makes sure that every intake question has been fully answered with nothing left open. In some cases there’s an added step: certain law firms want a background check on the claimant, usually looking at criminal history, so when a firm requires it we run one to their specifications. Others don’t ask for it at all, so it’s handled case by case depending on who we’re delivering to.

When something doesn’t clear: Not every case makes it through QA clean, and how we handle it depends on what’s wrong. If something is blatantly fabricated, a forged receipt or a story that falls apart, we disqualify it outright. If it’s less clear-cut, just a small doubt or a missing detail, the case goes back to a manager for a dedicated manager call, where they work directly with the claimant to fill the gap or clear up the inconsistency. Only once a case has cleared all of this does it move on to delivery.

Step 5: Delivery

The last step is getting the finished case into the client’s hands. Completed cases are transferred based on client specifications, typically via API integration into a case management system or through a secure shared folder. Because these files carry sensitive personal information, her ID, her contact details, and her account of what happened, we deliver them through secure, access-controlled methods rather than anything open. Technical integrations are configured upfront to ensure signed, QA-verified cases transfer immediately.

________________________________________________________________________________________________________________________________________________________________________________________