--- title: "How to Attribute Signups to PR Coverage Without Fooling Yourself" description: "How to attribute signups to PR coverage: what UTMs miss, how to read branded search lift and referral traffic, and why a source field beats dashboards." author: "Shilika Jain" date: "2026-10-28T02:07:50.255+00:00" tags: ["pr measurement", "pr roi", "earned media", "web3 pr", "ai search"] canonical: "https://www.shilikajain.com/blog/earned-media-attribution-user-growth" --- # How to Attribute Signups to PR Coverage Without Fooling Yourself By [Shilika Jain](https://www.shilikajain.com/authors/shilika-jain) - 10/28/2026 How to attribute signups to PR coverage: what UTMs miss, how to read branded search lift and referral traffic, and why a source field beats dashboards. --- # How to Attribute Signups to PR Coverage Without Fooling Yourself You can't attribute signups to PR coverage with the precision of a paid ad, so don't try. What works is a layered approach: track the referral traffic you can see, compare branded search and direct signups against a baseline before and after each story, ask new users how they heard about you, and treat the result as a confidence range rather than a single number. Combined, those signals tell you honestly whether coverage is moving growth. Founders usually ask me this right after a good piece runs. They open the analytics, see a modest spike in referral traffic, and wonder why the CoinDesk story they waited three months for "only" sent a few hundred visitors. The answer is that most of its effect doesn't arrive through the link. ## Why is PR so hard to attribute? Because the path from reading an article to signing up rarely runs through a click. A typical journey looks more like this: someone reads a story on their phone, doesn't click, sees the founder on a podcast a week later, searches the company name on a laptop, and signs up through the homepage. Your analytics records that as direct or organic search. PR gets no credit. A few technical reasons make it worse: - **Journalists don't use your UTM links.** Reporters link to your homepage or docs as they see fit, and most outlets won't add tracking parameters for you. - **Syndication scatters the traffic.** One story can be republished or summarised across aggregators, exchange news feeds and regional editions, each with its own referrer, or none. - **Dark social strips the trail.** When a link is shared in Telegram, Slack, WhatsApp or email, the referrer is often lost and the visit shows up as direct. - **The effect is delayed.** A story builds recognition that pays off when someone needs you weeks later. > classic problem of web3 startups: don’t understand CAC, LTV, turnover or other basic metrics! > > [Shilika Jain on X, 8 Dec 2022](https://x.com/Shilika_jain/status/1600724726149574656) That's still the trap. Teams that don't have basic acquisition metrics in place try to measure PR with the one tool they do have, last-click attribution, and conclude PR doesn't work. ## What can each attribution signal tell you? This is what I call the four-layer attribution stack. Each layer catches a different part of PR's effect, and each has a different level of confidence. You need all four, because no single one is reliable on its own. | Layer | What to track | What it catches | Confidence | What it misses | | --- | --- | --- | --- | --- | | 1. Direct referral | Sessions and signups with the outlet as referrer | Readers who clicked straight through | High, but small | Everyone who didn't click | | 2. Search and direct lift | Branded search impressions and clicks, direct traffic, against a baseline | People who heard of you and looked you up | Medium | Noise from other campaigns running at the same time | | 3. Self-reported source | A free-text "how did you hear about us" field at signup | Memory of where someone first noticed you | Medium, qualitative | Answers skew toward the last thing people remember | | 4. Downstream mentions | Sales calls, partner emails, community messages that cite coverage | Influence on higher-value conversions | Low volume, high signal | Anything nobody bothered to log | The top layer is the one dashboards show you. The bottom three are where most of the value sits. ## How do you set up attribution before a story runs? Most of this work has to happen before the coverage, not after. Once a story is live, you can't go back and create a baseline. My setup checklist for any announcement: 1. **Record a two-week baseline** of branded search impressions in Google Search Console, direct sessions, daily signups and organic signups. Note any other campaigns running. 2. **Add a free-text source field** to signup or onboarding. Make it optional but visible. Free text beats a dropdown, because people write "saw you in The Block" or "heard the founder on a podcast" in ways a list never captures. 3. **Create a clean landing page** for the announcement on your own site, so when reporters do link to it, those visits are easy to isolate. 4. **Prepare a short link** for your own social posts about the coverage, separate from whatever the outlet uses. 5. **Brief the sales and community teams** to log any mention of the story in calls, DMs or Telegram. 6. **Write down your expected range** before it runs. It stops you from rationalising any result afterwards. None of this takes long. It just has to be done the week before, not the morning after. ## How do you read the numbers after coverage lands? Look at a window, not a day. I compare the two weeks before with the two weeks after, and again with weeks three to six, because the delayed effect is often larger than the spike. An illustrative worked example, with made-up round numbers to show the method: say your baseline is 40 signups a day, with branded search impressions flat. A funding story runs in a major trade outlet. Over the next fourteen days, direct referral from the outlet accounts for a few dozen signups. Signups overall average 52 a day, and branded search impressions rise noticeably. Your source field shows a cluster of answers naming the outlet or "saw the funding news". What can you honestly say? Roughly a few dozen signups came directly from the article. Most of the remaining lift over baseline is plausibly linked to it, given the search lift and self-reported answers, unless something else changed that week. That gives you a range, from the direct number at the low end to the full lift above baseline at the high end. Report the range, not the top of it. > seems like announcing funding story is still a good user acquisition + social growth strategy! > > [Shilika Jain on X, 16 Oct 2024](https://x.com/Shilika_jain/status/1846622202072105403) Funding stories are a good test case because they often reach beyond trade readers. Measure one carefully and you'll have a reference point for every announcement after it. ## What about AI search and syndicated coverage? Two newer channels make attribution messier, and both are worth tracking anyway. **AI assistants.** More people now ask ChatGPT, Perplexity or Gemini which tools to use in a category. Those answers tend to cite the same outlets PR targets. Some assistant visits show up in analytics with an identifiable referrer, but many don't. The practical measure is to ask the major assistants your category questions once a month and note whether you're named, and what source they cite. I go into this in [how to get cited by ChatGPT](/playbook/how-to-get-cited-by-chatgpt-2026), and it's the core of my [AI search visibility work](/services/ai-seo). **Syndication.** When Tangem's self-custody survey ran as a data story, the same stat was picked up across Metaverse Post, Watcher.Guru, Binance Square, KuCoin News and Bloomingbit's English, Korean and Japanese editions, among others. Each of those has its own audience and referrer behaviour, and some regional readers will never click through to an English site at all. For coverage like that, count the pickups and markets reached, and watch for lift in those regions, rather than expecting the referral report to capture it. > googling best exchanges in crypto only shows you exchanges that have best SEO > > [Shilika Jain on X, 8 Dec 2022](https://x.com/Shilika_jain/status/1600830595600220160) Earned coverage also feeds that search layer. Links and mentions from credible outlets help determine what shows up when someone searches your category, which is another delayed effect no click report will credit. ## What should you not over-read? A few mistakes I see founders make once they start measuring: - **Treating a quiet referral report as failure.** A small click-through number from a respected outlet is normal. - **Crediting PR for everything in the window.** If an airdrop, a listing or a big KOL post ran the same week, the lift isn't PR's alone. - **Comparing PR cost per signup with paid ads.** They do different jobs. Paid buys a click now; PR builds the recognition that makes future clicks cheaper. - **Chasing a single attribution number for the board.** A confidence range with the method explained is more credible than a precise figure nobody believes. - **Ignoring downstream value.** One partnership conversation that started from coverage can matter more than hundreds of signups. The sibling post on [where PR shows up in a B2B pipeline](/blog/pr-and-sales-pipeline-realistic-expectations) covers that side. For how this fits into wider reporting, [how to measure PR results for startups](/blog/how-to-measure-pr-results-startups) and the [Web3 PR ROI framework for boards](/blog/web3-pr-roi-metrics-framework-founders-boards) go further. ## When is attribution not worth the effort? If you're doing one announcement a year, a full measurement setup is overkill. Add the source field, note the baseline and move on. And if your product has no self-serve signup, like an enterprise sale or a protocol that integrates through partners, signups are the wrong unit entirely. Track inbound conversations and partner mentions instead. Attribution also has a cost in attention. Teams that refresh the dashboard every hour after a story runs often stop doing the follow-up work, like sharing the piece with prospects and pitching the next angle, that actually extends the effect. ## FAQ ### Should I ask journalists to include UTM links? You can include a tracked link in your press materials, but don't expect reporters to use it, and don't push. Their job is to link where it's useful for readers. Put your effort into a clean landing page on your own site instead, so you can isolate visits however the link arrives. ### Is "how did you hear about us" data reliable? It's imperfect, because people remember the most recent or most memorable touch. But it's often the only signal that captures podcast listeners, word of mouth and offline reading. Read it in aggregate over weeks, and look for patterns rather than trusting any single answer. ### How long after a story should I keep measuring? At least six weeks for a significant piece. The initial spike fades within days, but branded search, inbound conversations and sales mentions can keep rising for weeks after, especially if you share the coverage actively. ### Can I attribute token holders or wallet activity to PR? Only very loosely. Onchain activity has too many drivers, from incentives to market moves, to isolate a single article's effect. If you want a signal, compare new wallet connections from your own site against your baseline, and treat any lift as a hint, not proof. *Want a measurement setup in place before your next announcement goes out? [Book a 30-minute teardown](https://calendly.com/shilikajain/30min/) and bring your current analytics.* --- **Book a 30-min PR measurement teardown with Shilika** - https://calendly.com/shilikajain/30min/ Canonical: https://www.shilikajain.com/blog/earned-media-attribution-user-growth