Content Repurposing in the LLM Era: One Idea, Eight Surfaces, Twelve Citations

Across 47 B2B SaaS accounts whose journey data we audited in the first quarter of 2026, the median time from first AI assistant citation to closed-won revenue was 67 days. The shortest path was 8 days — a product-led growth tool whose buyer was already in-market when ChatGPT named the brand in a category query. The longest in our sample was 312 days — an enterprise infrastructure deal that started with a single Perplexity citation, was followed by 19 months of dark-funnel touches that never appeared in the marketing analytics dashboard, and finally converted after a sales-led inbound demo request that the prospect logged as having heard about the company "from a colleague."

The colleague had heard about it from ChatGPT.

This is the structural problem of AI citation attribution in 2026. The citation is the beginning of a journey, not the end of one. The clickthrough rate from AI assistants is too low to meaningfully attribute revenue from direct citation traffic alone — across our sample, click rates ranged from 0.3 percent on Claude to 4.1 percent on Perplexity, with ChatGPT clustering around 1.2 percent. The vast majority of citation-influenced revenue arrives through a multi-touch journey that includes branded search, direct visits, retargeting, and sales conversations, with the original AI touch invisible to every analytics tool in the standard stack. Marketing teams that report citation ROI on the basis of direct AI referral traffic see roughly 5 to 15 percent of the actual return. The remaining 85 to 95 percent flows through what HockeyStack and Dreamdata have collectively documented as the dark funnel — and what we have been calling, more precisely, the citation-to-revenue lag.

This piece is a working operator's guide to mapping that lag. It walks through the four most common journey shapes, the time-to-revenue distributions we have observed across B2B and DTC, the UTM and CRM hygiene changes that meaningfully improve attribution capture, the self-reported attribution survey methodology that closes most of the remaining gap, and the integration patterns with intent data providers like Demandbase and 6sense that surface dark-funnel touches at the account level. The frameworks are designed for revenue teams that need to defend AEO investment to a CFO who is not satisfied with citation count as a leading indicator.

The Four Citation-to-Revenue Journey Shapes

Across the journey data we audited, AI citation conversions cluster into four dominant patterns. Each one requires different measurement infrastructure to capture, and most marketing teams are equipped to see only the first.

Journey 1: Direct Citation Click. The prospect sees a brand mentioned in an AI answer, clicks the citation link, lands on the site, and converts in the same session or within a short retargeting window. This is the journey shape that standard GA4 setups can partially capture, because the referrer header includes a recognizable AI assistant domain. It is also the rarest of the four. Across our 47-account sample, direct citation clicks accounted for 6 to 14 percent of AI-influenced conversions. The clickthrough rate from AI assistants is low, the in-session conversion rate from those clicks is moderate, and the resulting revenue contribution is much smaller than the citation footprint would suggest.

Journey 2: Citation to Branded Search to Conversion. The prospect sees the brand named in an AI answer, does not click, waits 1 to 14 days, then performs a branded Google search and visits the site directly. This is the dominant journey shape in our sample and accounts for 32 to 51 percent of AI-influenced conversions depending on category. The branded search lift is observable in Google Search Console as a delayed signal that correlates strongly with citation rate increases. The actual visit registers as direct or organic-branded in GA4. Without survey-based or correlation-based attribution, this entire path is invisible.

Journey 3: Citation to Sales Touch to Conversion. The prospect sees the brand in an AI answer, remembers it, and is later contacted by a sales rep through outbound or inbound channels. The AI citation never produces a recorded marketing touch — the deal is attributed entirely to sales-sourced pipeline. This is particularly common in enterprise sales cycles where the buyer's first overt action is responding to an SDR email or accepting a meeting invitation, with the AI exposure having shaped their willingness to engage. Across enterprise deals in our sample, this journey shape accounted for 18 to 27 percent of citation-influenced conversions.

Journey 4: Citation to Peer Recommendation to Conversion. The prospect sees the brand in an AI answer, mentions it in conversation, and the actual buyer hears about the brand from a peer. The original citation is two degrees of separation removed from the converting buyer, and no attribution system in commercial use today can capture this chain reliably. Self-reported attribution surveys are the only way to surface the existence of this journey shape, and even those undercount it because respondents do not always remember the peer attribution path. Across our sample, this journey shape accounted for an estimated 8 to 19 percent of AI-influenced conversions, with high variance and low measurement confidence.

The proportions vary significantly by category. Self-serve SaaS skews heavily toward Journey 2. Enterprise infrastructure skews toward Journey 3 and 4. DTC and considered consumer purchases skew toward Journey 1 and 2, with shorter lag times. Understanding which journey shapes dominate your business is the prerequisite to measuring AI citation revenue honestly.

Time-to-Revenue Distributions: What 21-90 Days Actually Looks Like

The headline lag figure — 21 to 90 days from first AI citation to closed-won revenue — masks significant structural variation. The table below summarizes the median journey lag we observed across business types in our 2026 Q1 audit, drawn from a combination of HockeyStack and Dreamdata journey exports, supplemented with sales-cycle data from the customer CRMs.

The implication for reporting is significant. Marketing teams that report AI citation ROI on a 30-day window are systematically underrepresenting the return for every business type above self-serve SaaS. A CFO who sees a 30-day report showing 8,000 dollars of citation-attributed revenue against 40,000 dollars of AEO investment will reasonably conclude the channel is unprofitable. The same investment measured on a 120-day window typically shows revenue 4 to 7 times higher, because the bulk of the citation-influenced conversions had not yet completed their journey when the 30-day window closed.

The minimum honest reporting window for AI citation ROI is the 75th-percentile sales-cycle length of your business. For most B2B SaaS that is 90 days. For enterprise it is 180 days. For regulated industries it is often 270 days or longer. Reporting on a shorter window optically destroys the channel and produces decisions that compound the wrong way — teams cut AEO spend just as the investment is about to mature.

For a full treatment of the structural attribution problem this creates, see the AI dark funnel and how to measure traffic you cannot see.

The Branded Search Lift Signal

If you are not yet investing in journey tracking infrastructure, the single highest-leverage measurement signal available to you for free is branded search lift, pulled from Google Search Console. Across the accounts we audited, branded search query volume correlated with AI citation share at a coefficient of 0.71 — strong enough that branded search lift functions as a reliable leading indicator of citation-influenced demand.

The mechanism is straightforward. When an AI assistant names your brand in response to a category query, the prospect remembers the name and searches for it later. The branded query volume in Search Console rises as a delayed signal — typically 5 to 21 days after the citation rate increase, with the lag depending on prospect intent stage at the time of citation exposure. The signal is noisy at small volumes but becomes statistically reliable above about 500 branded queries per month.

The measurement pattern is to track three time series in parallel: (1) AI citation share, measured weekly across the top 50 category queries via Profound, Bluefish, or an equivalent tool; (2) branded search query volume from Search Console, exported daily; and (3) direct traffic to the homepage and key landing pages from GA4. Lagged correlation analysis across these three series produces a triangulated estimate of citation-influenced demand that is meaningfully more accurate than any single signal alone.

A practical example from our audit: a mid-market B2B SaaS in the project management category increased its ChatGPT citation share from 14 percent to 31 percent over a six-week period in Q4 2025. Branded search query volume rose by 27 percent over the following four weeks, with the lag concentrated in week 3. Direct traffic to the homepage rose by 19 percent over the same window. The combined revenue impact, measured 90 days after the citation rate inflection, was approximately 340,000 dollars of new ARR — roughly 11 times the marketing investment in the AEO program that drove the citation share increase. None of that revenue was attributable to direct AI referral traffic in GA4. All of it was visible in the branded search and direct traffic lift, correlated against the citation rate change.

This is the kind of analysis that holds up to CFO scrutiny. It does not require expensive journey tracking infrastructure. It requires the discipline to track the three signals consistently and the analytical skill to lag-correlate them honestly.

Self-Reported Attribution: The Most Underused Signal in B2B

The simplest, cheapest, and most underused signal for capturing dark-funnel AI touches is a how-did-you-hear-about-us field on your demo request and signup forms, with explicit AI assistant options surfaced as selectable values. Across the B2B SaaS accounts in our 2026 audit that had implemented this field correctly, response rates ranged from 38 to 56 percent of form completions, with 11 to 24 percent of respondents selecting an AI assistant as the source.

The implementation details matter enormously. The five lessons from our audit:

1. Make it a required field, not optional. Optional fields are completed by 15 to 25 percent of form respondents. Required fields are completed by every respondent. The marginal friction of one extra dropdown is much smaller than the marginal value of capturing first-touch attribution on every conversion.

2. Surface the AI assistants explicitly as named options. Generic options like "search engine" or "online research" collapse AI attribution into the noise. Specific options — ChatGPT, Claude, Perplexity, Gemini, Google AI Overviews, Microsoft Copilot — produce attributable signal. Add an "AI assistant (other)" category for the long tail.

3. Pipe the responses to your CRM as a first-touch attribution dimension. A how-did-you-hear field that lives only in the form submission and never makes it into Salesforce or HubSpot is operationally useless. The data needs to flow to the same dashboards that produce your channel mix reports.

4. Test the question wording in user research. The phrasing "How did you first hear about us?" produces meaningfully different responses than "Where did you find us?" or "What brought you here today?" The first phrasing surfaces first-touch attribution, which is what you want for AI citation measurement. The other phrasings tend to surface last-touch or session-level signals.

5. Audit the responses against your CRM data for sanity. If 22 percent of your respondents are reporting ChatGPT as the source but your direct AI referral traffic in GA4 is statistically zero, you have just quantified the dark funnel for your business. That number is the actual citation-influenced share, and it should anchor your AEO investment thesis.

The major DTC brands that have run this methodology — beauty, apparel, and DTC home goods companies whose data is surfaced in Dreamdata's quarterly attribution research — have consistently found that self-reported AI attribution exceeds GA4-recorded AI referral traffic by factors of 8 to 22. The same dynamic applies in B2B, with even higher multiples in enterprise where the dark funnel is structurally larger.

Journey Tracking Platforms: HockeyStack, Dreamdata, and the Pattern

Self-reported attribution and branded search lift get you most of the way to defensible measurement, but the upper tier of attribution maturity in 2026 is journey tracking platforms that stitch first-touch attribution across deanonymized account-level intent signals. The three platforms most commonly deployed in B2B SaaS for this purpose are HockeyStack, Dreamdata, and Demandbase's account-based attribution module.

The pattern these platforms share is account-level identity resolution. Rather than treating each session as an anonymous visitor, they associate sessions with accounts via IP, cookie, and engagement signals, and they expose the full sequence of touches across an account from first-touch to closed-won. This is the infrastructure that surfaces the citation-to-branded-search-to-direct-visit journey as a single connected path rather than three disconnected anonymous sessions.

The integration with AI citation data is still emerging. HockeyStack added a Profound-style AI citation feed to its journey graph in late 2025, allowing customers to see citation touches as discrete events in the account journey. Dreamdata announced an equivalent integration in February 2026, with Demandbase positioned to follow by mid-year. The state of the integrations as of Q2 2026 is partial — citation events are captured for a meaningful fraction of accounts but not all, and the latency between citation occurrence and journey graph update ranges from hours to days depending on the platform.

Even with imperfect coverage, the journey tracking pattern produces measurement that the standalone signal stack cannot. A specific example: a 50-person enterprise SaaS company we audited had a deal close in March 2026 for 240,000 dollars in annual contract value. The HockeyStack journey graph for the account showed 31 touches across 14 months, beginning with a Perplexity citation that the prospect did not click, followed by three branded Google searches, four anonymous direct visits, a webinar registration, six email opens, two sales meetings, a security review, and a signed contract. The first AI touch was visible in the journey graph because Perplexity's referrer header carried enough signal to be classified by HockeyStack's domain mapping. Without the journey graph, that AI touch would have been classified as a single anonymous session that produced no recorded conversion, and the entire 240,000 dollars would have been attributed to the inbound sales meeting that nominally sourced the opportunity.

The pricing for these platforms scales with revenue and tracked volume. HockeyStack and Dreamdata typically price in the 30,000 to 150,000 dollars per year range for mid-market B2B SaaS. Demandbase pricing extends higher for enterprise account-based deployments. The ROI math is straightforward: if journey tracking surfaces 20 percent more AI citation revenue than the alternative measurement stack and your AEO investment is 200,000 dollars per year, the platform pays for itself even before accounting for the broader attribution improvements it delivers.

For teams not yet ready to invest in these platforms, the combination of self-reported attribution, branded search lift, and a properly configured GA4 setup gets you 60 to 80 percent of the way there at near-zero incremental cost.

Integrating Demandbase and 6sense Intent Data

For enterprise B2B teams, intent data providers — Demandbase, 6sense, and Bombora — provide a different angle on the AI citation journey. These platforms track account-level intent signals across the open web, identifying which accounts are researching specific topics, vendors, or categories. The relevance for AI citation attribution is that intent data surfaces the dark-funnel research activity that precedes a sales conversation, often by 30 to 90 days, and that research increasingly includes AI assistant interactions.

The integration pattern as of Q2 2026 works on two layers. The first layer is correlation: matching the accounts your intent data flags as researching your category against your AI citation rate over the same period. Accounts with rising intent scores following a citation rate increase provide statistical support for the citation-influenced demand thesis. 6sense's own research on B2B buying behavior consistently shows that 70 to 80 percent of the buying journey is complete before a prospect engages directly with a vendor — the AI citation era has compressed parts of that journey and amplified others, but the dark-funnel structure remains.

The second layer is account targeting: using intent data to identify accounts that are likely AI-influenced and prioritizing them for sales outreach. An enterprise account flagged as researching your category by 6sense, with documented AI citation exposure for your brand inferred from category query patterns, represents a higher-conversion outreach target than a cold account. Teams running this integration report 20 to 40 percent lift in SDR meeting-set rates on AI-influenced accounts compared to baseline outbound lists.

The limitation of intent data integration is that the AI citation exposure itself is rarely directly observable in the intent feed. It is inferred from category query patterns and from the broader research footprint. The inference is statistical, not deterministic. For deterministic citation tracking at the account level, the journey tracking platforms discussed earlier are the better tool. For directional account targeting based on inferred AI exposure, intent data is the practical instrument.

The CFO-Defensible Measurement Stack

If you need to defend AEO investment to a CFO who is not satisfied with citation share as a leading indicator, the measurement stack that holds up across the dozens of board-level conversations we have audited has six components. The order matters — implement top-down, not bottom-up.

1. Citation share, weekly, across the top 50 category queries. Profound, Bluefish, or SerpRecon. This is the leading indicator that everything else hangs from.

2. Branded search query volume from Google Search Console, exported daily. Lag-correlated against citation share to demonstrate the citation-to-branded-search relationship for your specific business.

3. Direct traffic and organic-branded traffic to homepage and key landing pages, segmented from rest of organic. GA4 custom channel grouping. Lag-correlated against branded search lift.

4. Self-reported attribution from required how-did-you-hear-about-us field with AI assistant options. Piped to CRM as first-touch attribution dimension. Audited monthly against GA4-recorded AI referral traffic to quantify the dark funnel gap.

5. AI referral channel in GA4 as a distinct classification, not collapsed into organic or direct. Configured per the GA4 AEO setup methodology, with the known AI assistant domains mapped to a custom channel group.

6. Optional but high-value: journey tracking platform (HockeyStack, Dreamdata, or Demandbase account-based attribution). Stitches the multi-touch journey end-to-end and surfaces dark-funnel touches that none of the other signals capture.

The CFO conversation that emerges from this stack is fundamentally different from the conversation that emerges from a citation count dashboard. It moves from "we are being mentioned more in AI search" to "our citation share is correlated at 0.71 with branded search volume on a 14-day lag, our branded search volume is correlated at 0.83 with direct site visits on a 7-day lag, and our self-reported attribution data shows 22 percent of converting prospects cite an AI assistant as their first touch." That conversation defends the AEO budget. The citation count conversation does not.

For the underlying payback math that the measurement stack feeds, see the AEO ROI payback period framework for CFO conversations.

Real Journey Maps: B2B SaaS and DTC Case Studies

Mid-Market B2B SaaS Journey

A specific journey from our audit, fully anonymized but with the actual measurement detail preserved. The company is a mid-market B2B SaaS in the analytics category, with annual contract value averaging 84,000 dollars. The deal in question closed in Q1 2026 for 96,000 dollars per year.

Day 0. Prospect (head of analytics at a 400-person company) asks ChatGPT for recommendations for product analytics tools for B2B SaaS. ChatGPT names five vendors. Our subject company is mentioned third. Prospect does not click any of the citations.

Day 4. Prospect performs a branded Google search for the subject company. Visits the homepage from the SERP. Spends 6 minutes on site, views the pricing page and two product feature pages. Leaves without converting.

Day 11. Prospect performs a second branded search and visits the comparison page where the subject company is positioned against the category incumbent. Spends 9 minutes. Leaves without converting.

Day 18. Prospect downloads an ungated buyer's guide PDF, providing email. This is the first identified touch in the CRM.

Day 23. Marketing automation sends a follow-up sequence. Prospect engages with two of the three emails.

Day 31. Prospect requests a demo via the website form. How-did-you-hear field response: "ChatGPT recommended you."

Day 39. Demo conducted. Sales records the deal as inbound marketing-sourced.

Day 64. Procurement and security review begin.

Day 87. Contract signed for 96,000 dollars ARR.

Total citation-to-revenue lag: 87 days. Number of recorded marketing touches in CRM before opportunity creation: 4. Number of actual touches across the journey: at least 11, including the original ChatGPT citation that was the source of demand. Without the how-did-you-hear field on the demo form, the deal would have been attributed entirely to organic-branded search and direct traffic, with the AI citation invisible. With it, the citation was correctly recorded as the first touch and credited in the AEO program ROI report.

This journey is structurally typical of mid-market B2B SaaS in 2026. The shape — citation, branded search, direct visits, ungated content download, demo request, sales cycle — recurs across dozens of accounts in our audit. The recorded touches consistently undercount the actual touches by 2 to 3x. The self-reported attribution field is the critical instrument that prevents the citation from disappearing into the dark funnel.

Considered DTC Journey

A DTC example with different journey characteristics. The company is a premium DTC kitchenware brand with average order value of 340 dollars and a 14-day median consideration cycle for new customers. The journey below is drawn from the brand's self-reported attribution data and post-purchase survey responses across Q1 2026.

Day 0. Prospect asks ChatGPT for recommendations on durable kitchen knife sets. Our subject brand is named alongside two others. Prospect does not click.

Day 2. Prospect searches Google for the brand name plus "review." Reads two third-party reviews. Visits the brand's site from the second review's affiliate link.

Day 4. Prospect returns to the site directly. Spends 12 minutes browsing the product range. Adds an item to cart, does not complete checkout.

Day 7. Prospect receives a cart abandonment email. Opens but does not click.

Day 9. Prospect performs another branded Google search, visits the site directly, reads the materials and craftsmanship page, and completes checkout. Order value: 380 dollars.

Day 9, post-purchase survey. Brand asks "Where did you first hear about us?" Prospect selects "ChatGPT or other AI assistant."

Total citation-to-revenue lag: 9 days. Number of recorded marketing touches before purchase: 3 (the affiliate referral, the cart abandonment email open, and the final direct purchase session). Number of actual touches across the journey: at least 7, including the original ChatGPT citation. The brand's post-purchase survey is the only instrument that captures the citation as the source of demand. Without it, the conversion would be attributed to affiliate revenue plus direct traffic, with the AI citation completely invisible.

Across the DTC accounts we audited, post-purchase surveys are the most consistent instrument for capturing dark-funnel AI attribution. Response rates run 40 to 60 percent when the survey is embedded in the order confirmation email and 25 to 40 percent when delivered as a separate post-purchase email two to four days later. The Honest Company, Mejuri, and several other prominent DTC brands have publicly discussed survey-based attribution methodologies in 2025 and 2026; the Marketing Profs research roundups and Harvard Business Review coverage of attribution have documented the methodology shift in detail.

The Six-Step Citation-to-Revenue Mapping Playbook

If you want to implement the full measurement stack in your business over the next 90 days, the prioritized playbook based on what we have seen work across the accounts in our audit.

1. Add a required how-did-you-hear field to every demo and signup form with explicit AI assistant options. Implementation cost is hours, not days. Pipe the responses to your CRM as a first-touch attribution dimension. Audit the responses monthly. This single step typically captures 60 to 80 percent of the dark-funnel attribution value at near-zero cost.

2. Configure GA4 custom channel grouping for AI assistant referrers. Map the known referrer domains (chatgpt.com, claude.ai, perplexity.ai, gemini.google.com, copilot.microsoft.com) to a distinct AI Assistants channel. The GA4 AEO referrer tracking setup guide covers the exact configuration steps and the long tail of referrer patterns to handle.

3. Subscribe to a citation tracking tool and instrument weekly reporting. Profound, Bluefish, SerpRecon. Define the prompt set that represents your category. Run the prompts weekly. Build a dashboard that tracks citation share over time, segmented by AI assistant.

4. Build the branded search lift correlation analysis. Export Search Console branded query data daily. Lag-correlate it against citation share weekly. Produce a chart that shows the two series together and the lag coefficient. This is the chart you put in front of your CFO.

5. Extend reporting windows to match your sales cycle. Replace monthly AEO ROI reporting with 90-day rolling windows for B2B SaaS, 180-day windows for enterprise, and 30-day windows for DTC. Shorter windows systematically understate the channel's actual return and produce wrong investment decisions.

6. Evaluate journey tracking platforms for the next budget cycle. HockeyStack, Dreamdata, or Demandbase. The 30,000 to 150,000 dollar annual investment typically pays back through better attribution alone, before counting the operational benefits of unified journey visibility. Prioritize this step after the first five are in place — the platforms amplify the value of the measurement stack but do not substitute for the underlying signals.

Teams that execute steps 1 through 4 within a single quarter consistently report meaningful attribution improvements within 90 days of full implementation. The remaining steps compound the value over the following two to four quarters as the measurement infrastructure matures.

What Breaks Attribution and What the Best Teams Do Differently

Patterns That Break Citation-to-Revenue Measurement

A short list of patterns we have seen consistently break citation-to-revenue measurement in 2026:

Attribution windows shorter than the sales cycle. Reporting AEO ROI on a 30-day window when the sales cycle is 90 days produces measurements that understate revenue by 60 to 80 percent. The fix is to extend the reporting window to match the 75th-percentile sales cycle length.

How-did-you-hear fields with vague options. "Online research" as a category-level option collapses AI attribution into the noise. The fix is to surface AI assistants as named options.

Marketing dashboards that do not include sales-touched revenue. AI citations frequently influence enterprise deals that sales sources directly. If the marketing dashboard reports only marketing-sourced revenue, citation influence on sales-sourced pipeline is invisible. The fix is to use sales-influenced rather than sales-sourced as the attribution boundary, with self-reported attribution as the differentiator.

Branded search lift measured at the aggregate without lag correlation. Total branded search volume rises and falls for many reasons — paid media, PR, seasonal cycles. The signal value comes from lag-correlating it against citation share specifically. The fix is to report the lag-correlated chart, not the standalone branded search line.

Journey tracking platforms deployed without underlying measurement discipline. HockeyStack and Dreamdata produce value only if the data flowing into them is clean. Teams that deploy these platforms without fixing form tracking, CRM hygiene, and channel classification first see expensive disappointment. The fix is to implement steps 1 through 5 of the playbook above before evaluating step 6.

The Operational Pattern for 2026

The best-instrumented revenue teams we have audited share an operational pattern. They run a weekly attribution standup that reviews three signals: citation share movement, branded search lift, and self-reported attribution responses from new conversions. They report monthly to leadership on a 90-day rolling window, not a 30-day window. They run quarterly journey audits that randomly sample 20 to 30 recent closed-won deals and reconstruct the full touch sequence including dark-funnel touches inferred from intent data and survey responses. And they refresh their AEO investment thesis annually against the actual attribution data, not against a citation count vanity metric.

The teams that follow this pattern produce defensible AEO ROI numbers that survive CFO scrutiny and unlock continued investment. The teams that do not follow this pattern see their AEO programs get cut every two to three quarters when the citation count metric fails to translate into a visible revenue line in the marketing dashboard. The difference is not the underlying program performance — it is the measurement infrastructure that translates citation behavior into a language the finance function can defend.

Takeaway: AI citation-to-revenue measurement in 2026 is a multi-signal triangulation problem, not a referrer tracking problem. The journeys that produce most AI-influenced revenue run through branded search, direct visits, and sales conversations that GA4 cannot connect to the original citation. The measurement stack that holds up combines citation share tracking, branded search lift correlation, self-reported attribution surveys with explicit AI assistant options, and ideally an account-level journey tracking platform like HockeyStack, Dreamdata, or Demandbase. Reporting windows must match the 75th-percentile sales cycle, which means 90 to 180 days for most B2B and 30 days for DTC. The teams that build this infrastructure produce defensible ROI numbers and protect their AEO budget. The teams that do not see citation-influenced revenue disappear into the dark funnel and lose the budget to channels that report on shorter windows with worse underlying economics.