Build a Unified Intent Graph Across YouTube, TikTok

Seventy-three percent of B2B buyers now touch five or more discovery channels before a vendor even knows they exist, according to Forrester research. YouTube search. TikTok discovery feeds. Instagram Explore. ChatGPT and Perplexity conversations. Each channel generates intent signals, and almost none of them talk to each other. The result? Agencies and growth teams spray budget across platforms, prospecting the same buyer in five different places — or worse, missing them entirely. The fix isn’t another dashboard. It’s a unified intent graph that stitches fragmented discovery signals into a single, actionable map.

Why Discovery Fragmentation Broke the Old Funnel

The linear funnel assumed buyers searched Google, clicked an ad, visited a landing page, and converted. That model was already cracking by 2023. It’s now fully shattered. A CMO researching project management tools might watch a YouTube comparison video at 7 AM, scroll past a TikTok creator review during lunch, ask ChatGPT for a shortlist at 3 PM, and browse Instagram Explore ads that evening. Each of those moments carries intent data. None of it lands in the same system by default.

This isn’t just a measurement problem. It’s a spending problem. When media buyers can’t see that the same prospect just expressed intent on three platforms, they bid on that prospect three separate times — inflating CAC and cannibalizing their own pipeline. Agencies running predictive budget reallocation models still struggle when the underlying signal layer is fragmented.

The concept of a unified intent graph isn’t theoretical anymore. It’s an architectural requirement.

What a Cross-Platform Intent Graph Actually Looks Like

Think of an intent graph as a living data structure where nodes represent identities (people, accounts, device clusters) and edges represent intent signals weighted by recency, depth, and platform context. Unlike a CDP, which primarily stores profile attributes and event histories, an intent graph prioritizes relationships between signals over static records.

Here’s the architecture at a high level:

The unified intent graph doesn’t replace your CDP or analytics platform. It sits between raw platform data and activation, serving as the intelligence layer that tells you who wants what and how badly — regardless of where they showed up.

The Identity Resolution Problem Nobody Wants to Talk About

Identity resolution across walled gardens is brutally hard. Meta, Google, and TikTok have spent billions making sure their user data stays inside their ecosystems. You’re not getting deterministic cross-platform matching from their APIs. Period.

So what actually works?

First-party data anchors. If you have email or phone from a gated asset, webinar registration, or CRM record, that becomes your deterministic spine. Every probabilistic signal — IP-based clustering, browser fingerprinting within privacy bounds, temporal pattern matching — attaches to that spine. Companies like HubSpot and Clearbit have made strides here, but they still operate primarily within web and email contexts.

The real unlock for intent graphs is behavioral fingerprinting: identifying that a user who searched “enterprise data platform comparison” on YouTube at 9:47 AM from a corporate IP range is probabilistically the same person who asked Perplexity “top enterprise data platforms for mid-market” twelve minutes later from the same geographic cluster. You won’t get 100% match rates. You don’t need them. At 60-70% confidence with proper decay weighting, the graph already outperforms siloed platform signals by a wide margin.

Privacy compliance is non-negotiable here. Your identity resolution must respect GDPR, CCPA, and the evolving patchwork of state-level privacy laws. The graph should store probabilistic match scores, not PII, at the edge level — keeping the architecture auditable. For teams navigating complex compliance stacks, understanding your data privacy obligations is table stakes.

Signal Weighting by Platform: Not All Intent Is Created Equal

A TikTok like is not the same as a YouTube search query. An Instagram Explore tap is not equivalent to a ChatGPT conversation where someone asks “Which tool should I buy for X?” Yet most attribution models treat cross-platform engagement as roughly interchangeable. That’s lazy, and it costs money.

Here’s how signal weighting should differ by platform:

• YouTube Search: High intent. Users actively type queries. Watch time past 50% on comparison or review content is a strong buy signal. Weight: 0.8–0.9 on a normalized scale.
• TikTok Discovery: Lower individual intent but high volume. Saves and shares indicate more intent than likes. Following a creator in a product category adds signal. Weight: 0.3–0.5, boosted to 0.6+ when combined with saves or comment engagement.
• Instagram Explore: Mid-range. Explore taps show curiosity; profile visits and story interactions suggest deeper interest. Weight: 0.4–0.6.
• AI Chat Assistants (ChatGPT, Perplexity, Gemini): Extremely high intent. Conversational queries are specific, often transactional. “Which CRM integrates best with Salesforce under $50/seat?” is a buyer talking. Weight: 0.9–1.0. The challenge is capturing these signals, which often requires integration with your own AI-surfaced content or referral tracking.
• Conversational Search (Google AI Overviews, Bing Copilot): High intent, similar to traditional search but with less click-through data available. Weight: 0.7–0.85, adjusted by query specificity.

These weights aren’t static. They should decay over time (a YouTube search from 48 hours ago matters more than one from three weeks ago) and compound when signals stack across platforms. Someone who searched YouTube, saved a TikTok, and asked ChatGPT about the same category within a 72-hour window? That’s a red-hot lead regardless of which single platform score looks highest.

Teams already leveraging cultural signal mapping will recognize this pattern — the value isn’t in any single signal but in the velocity and convergence of multiple signals pointing at the same intent.

Eliminating Redundant Prospecting Spend

Here’s where the unified intent graph pays for itself in weeks, not quarters.

Consider a mid-market SaaS company spending $200K/month across YouTube pre-roll, TikTok Spark Ads, Instagram Reels placements, and sponsored AI chat results. Without a unified graph, their frequency capping is platform-siloed. A prospect sees the brand seven times on YouTube, four times on Instagram, twice in TikTok, and once via a sponsored Perplexity result. That’s fourteen impressions. Maybe six of them were redundant. At a $12 average CPM blended across platforms, that’s real money burning — and it compounds across thousands of prospects monthly.

With a graph in place, media buyers can implement cross-platform frequency management. Once a prospect reaches a threshold intent score via any combination of platforms, the system suppresses prospecting ads and shifts budget toward conversion-stage messaging — or reallocates entirely to new audiences. This is the same principle behind predictive media buying strategies, but applied at the identity level rather than the campaign level.

Agencies get an additional benefit: client reporting that actually shows deduplicated reach. When you can demonstrate that you influenced 8,000 unique accounts across five platforms (rather than claiming 25,000 “touches” that are mostly the same 4,000 people), your strategic credibility goes through the roof.

Prioritizing the Highest-Intent Touchpoints

The graph doesn’t just eliminate waste. It reranks your entire prospecting queue.

Instead of platform-specific lead lists — “here are our hottest YouTube viewers, here are our most engaged TikTok followers” — the graph produces a single ranked list of accounts and individuals sorted by composite intent score. A prospect who barely registered on any individual platform’s radar might rank in your top 50 because their cross-platform signal pattern shows clear buying behavior.

For sales teams, this changes outreach sequencing. For media teams, it changes bid strategies. For agency strategists, it changes how they allocate creative resources — you build conversion assets for the platforms where your highest-intent cohort actually is, not where your biggest budget historically sat. That kind of insight pairs well with approaches like real-time cultural moment scoring, where timing and context amplify an already-strong signal.

The shift is fundamental: stop organizing strategy by platform, start organizing strategy by intent.

Where to Start

You don’t need to boil the ocean. Start with two platforms where you have the richest first-party data overlap (typically YouTube and one social channel), build identity resolution against your CRM spine, and prove the concept by measuring deduplicated reach against one campaign. The architecture scales from there. The agencies that build this capability now won’t just save budget — they’ll see buyers their competitors literally cannot.

FAQs

What is a unified intent graph?

A unified intent graph is a data architecture that stitches together buyer intent signals from multiple discovery platforms — YouTube, TikTok, Instagram, AI chat assistants, and conversational search tools — into a single, identity-resolved map. It enables teams to see composite buying signals across channels rather than relying on siloed, platform-specific engagement data.

How does identity resolution work across walled gardens like Meta and Google?

Deterministic matching uses known identifiers like email or phone from your CRM as an anchor. Probabilistic matching then attaches behavioral signals — such as temporal query patterns, IP clustering, and geographic data — to those anchors. Match rates of 60-70% are typically sufficient to outperform siloed platform signals significantly.

Why should intent signals be weighted differently by platform?

Different platforms reflect different levels of buyer intent. A YouTube search query indicates active research (high intent), while a TikTok like reflects passive discovery (lower intent). Weighting signals by platform ensures your scoring model accurately reflects actual purchase readiness rather than treating all engagement equally.

How does a unified intent graph reduce wasted ad spend?

By resolving buyer identities across platforms, the graph enables cross-platform frequency management. Once a prospect reaches a target intent score, prospecting ads are suppressed and budgets shift to conversion messaging or net-new audiences. This typically eliminates 25-35% of redundant impression spend.

Can smaller agencies or teams implement a cross-platform intent graph?

Yes. Start with two platforms where you have strong first-party data overlap, build identity resolution against your existing CRM records, and validate the approach on a single campaign. Graph databases like Neo4j offer free tiers, and the architecture scales incrementally as you add platforms and signal sources.