Top of funnel
Reaching people who do not know you yet.
Nine programs for reaching people who do not know you yet, and the discipline that keeps a touch's real effect properly credited to the channel that actually produced it.
Top of funnel is every program aimed at someone who has not yet registered that a business exists, or has not consciously connected a name to a need. Nine programs run here — paid search, paid social, SEO, generative engine optimization, answer engine optimization, programmatic advertising, creative ads strategy, campaign launch strategy, and influencer marketing — covering the full range of platforms and formats a first exposure can happen on.
What makes measurement hard here
The defining problem at this stage is displacement, not absence. A touch at this stage almost always has a real effect, but the evidence of that effect routinely shows up somewhere other than inside the touch's own report. Someone sees a paid social ad, a programmatic placement, an influencer's post, or a search result, remembers the name, and searches for it later — often through branded search, which gets credited entirely to that later, cheaper channel. A view-through window, a rank position, or an impression count is a real signal, but it answers whether someone was exposed to something, a different question entirely from whether that exposure caused a customer, and every program on this page has some version of that gap.
The fix that recurs across almost all nine programs is the same shape even where the mechanics differ: a matched-market or geographic holdout that compares a market where the program ran against a comparable market where it did not, paired with branded-search-lift tracking that looks specifically for the delayed, displaced signal a click-based report would otherwise miss. Neither tool is specific to one channel here — they are the answer to what top-of-funnel measurement actually requires.
This is also why a report built purely from platform dashboards tends to flatter whichever program happens to sit closest to the eventual click, and quietly undercount everything upstream of it. A campaign that introduced a business to a hundred people who searched the brand a week later looks, in a click-based report, exactly like a campaign that reached no one at all. Correcting for that is a load-bearing part of the reporting — without it, budget drifts toward the channel nearest the finish line and away from whatever is actually filling the funnel.
How these programs combine
Paid search and SEO typically split by how much organic visibility a business already has for its non-brand terms; where SEO is still building, paid search covers the gap and hands off as rankings improve. Programmatic supplies the volume display, CTV, and audio buying needs at scale once a concept is validated, while creative ads strategy is what validates the concept in the first place — the two rarely run independently of each other. GEO, AEO, and SEO together cover the full surface a search engine or an AI assistant can show a business on, each with its own mechanism but a shared reliance on the same well-structured content. Influencer marketing sits alongside paid social as a way to borrow an audience someone else already built, and campaign launch strategy runs underneath any of these specifically during the ninety days a new location or service line has no baseline yet to measure against.
Where to start
An operator with an established brand and some existing organic visibility usually starts with SEO or the GEO/AEO pair, depending on whether the more pressing gap is conventional search or AI assistant visibility, since both compound in value the longer they run and are the cheapest programs on this page over time. An operator opening a new location or entering a market for the first time starts with campaign launch strategy specifically, since every other program on this page assumes a baseline already exists to measure against, and that one is built for the ninety days before it does.
An operator with paid budget already committed but no real read on what it produces starts with whichever paid program is receiving the largest share of that budget today, run through a holdout before any reallocation decision gets made. Chasing a bigger number without first confirming the current one is real is the most common way top-of-funnel budget gets spent twice: once on the program, and again correcting for a decision made on a number that was never trustworthy.
Every program here ultimately reports through the same attribution layer that connects it to a booked outcome, and the programs that convert this stage's demand into someone who has decided but not yet purchased are covered at middle of funnel.
Paid social prospecting on Meta, LinkedIn, and TikTok measured through geo holdouts and matched-market tests rather than the platform's own view-through conversion count, which credits impressions nobody consciously noticed. Covers creative volume as an ongoing dependency, audience structure that excludes existing customers, and platform allocation built from where a specific audience actually spends attention rather than a fixed default split.
Non-brand paid search managed to cost per booked outcome. Covers Google Ads and Local Services Ads with call scoring to remove unqualified and duplicate calls, weekly search term hygiene against brand and competitor drift, and budget set and moved per location based on actual capacity rather than a blended regional average. LSA's own reporting currency is reconciled against your call log, since its dashboard cannot see whether a lead actually booked.
Ad creative concepted and tested as falsifiable hypotheses rather than design variations, with a testing structure built specifically to survive platform auto-optimization, which otherwise reallocates budget toward an early leader before a test reaches statistical significance. Covers hypothesis framing, even-split enforcement, and pre-committed sample sizes calculated from real account traffic. Producing the resulting assets at volume, once a hypothesis is validated, is a separate, related discipline with its own page.
Launch planning for a new location or service line, where the defining problem is that there is no history yet to measure against. Instrumentation goes in before spend does, and the first ninety days are spent establishing that baseline. Structured around three decision gates — weeks four, eight, and twelve — each with criteria agreed before launch for continuing, changing, or stopping a specific channel or the launch as a whole.
Non-brand organic acquisition and the technical work underneath it, measured with a scan grid per location rather than one national rank number. Covers technical fixes, duplicate location page rebuilds, and content built against real non-brand query gaps, with honest timelines that separate fast technical wins from slower authority-driven gains. Branded and local capture and AI assistant visibility are separate, related disciplines with their own pages.
Visibility inside AI assistant answers, measured with a fixed baseline question set run repeatedly against the assistants a business's buyers actually use, since there is no rank position to track. Covers source auditing, content restructured for direct extraction, and honest treatment of what cannot be promised — no vendor can guarantee a specific model recommends a business, and this page says so plainly rather than softening it.
Featured snippets, People Also Ask boxes, and voice results, measured as their own channel rather than folded into standard organic rankings. Covers the specific problem this discipline runs into that most others do not: winning a snippet can measurably reduce click-through to the page that won it, so presence has to be tracked and reported as its own metric, with an honest account of when zero-click visibility is worth having anyway.
Programmatic display, CTV, and audio bought with inventory identified placement by placement, fraud and viewability filtered before spend is counted, and the view-through attribution window shortened to a length that cannot claim credit for nearly anything that happens afterward. Distinct from retargeting, which defines an audience rather than a place an ad runs — the two overlap only where a retargeting campaign happens to be bought programmatically.
Creator partnerships measured with matched-market testing, since a media kit's impressions and follower counts can both be purchased and neither one is a measurement. Tracked codes and links undercount a creator's real effect badly, since the common response to seeing a creator's post is remembering the brand and searching for it later, which lands as branded search credited to that channel.
Find out which of your top-of-funnel spend is actually reaching new demand.
A review of your current top-of-funnel programs against branded search volume, showing how much of your reported activity is really discovery versus a later, displaced touch.