Client Reporting Beyond Clicks: Build the Three-Layer View

Client reporting that still opens with sessions is leading with the one layer that has gone blind, because only about a third of searches now produce any click, according to SparkToro. The fix is a three-layer record of exposure, behaviour and business outcome, with at least one layer sourced outside the platforms being reported on. Baselines cannot be backdated, so start this quarter.

The story going round agency floors is that attribution, and with it client reporting, got harder this year. It did not get harder. For most searches it stopped existing, which is a different problem with a different fix.

Your client reporting probably still opens on sessions, then sources, then a conversion number with a footnote. That page once stood in for the whole journey. It now covers the only stretch the platforms can still see. The replacement is three layers on one page, and one decision inside it that cannot wait.

Only About a Third of Searches Still Produce a Click

Clicking has stopped being the default outcome of a search. So far in 2026, only about a third of searches produce any kind of click, per SparkToro's measurement as reported in Search Engine Land's 2027 planning analysis. Globally, DataReportal's Digital 2026 overview puts search-engine use at its lowest in GWI's survey series, 80.3% of online adults in the past month.

Has it reached Manila? Two local numbers point opposite ways. Filipino AI-assistant usage ranks among the world's highest, at 42.4% of internet users aged 16 and over using ChatGPT monthly against a 26.5% global average, as PhilstarLIFE reported. The fieldwork is GWI's, from Q2 2025, published in the Digital 2026 report by Meltwater and We Are Social, and Meltwater competes with Media Meter. Referrals say otherwise. Statcounter's tracking-network data put AI chatbots at 0.8% of Philippine referral traffic in September 2026, against 81.98% for search.

Filipinos use the assistants heavily and arrive from them almost never, which is the argument for giving exposure its own layer in client reporting. Local vendors already sell against the gap: a Greenpark panel on PH AI visibility with Unilab, PepsiCo and Nestlé Nutrition digital leads put nearly one in three Philippine searches as already returning an AI answer first, and prescribed an audit tracking share of voice and a category's top cited sources. Treat that as vendor-led panel material with no published method, and its 60% figure for the next 12 to 18 months as a projection.

The Direct-Traffic Column Is Where AI Influence Hides

Influence created inside an AI answer still lands on your client's site, wearing no label. SparkToro and Similarweb tracked clickstream behaviour for seven days after AI recommended Capital One: direct visits to those brands rose while search visits ran about 15% lower, with the lift reaching 14.2%. A client reporting dashboard filing direct traffic as "unknown" is filing AI influence as noise.

Presence also has to be won surface by surface, because it does not travel. The same analysis found 91% of AI citations appearing in only one of ChatGPT, Perplexity or AI Overviews, and third-party signals driving 85% of brand visibility in large language models, per AirOps. Most work that moves how a brand shows up in AI answers happens where none of your tags are installed.

That is already the Philippine default. At a MARKETECH APAC session on budol culture, Primer Group's Tiffany Batungbacal reported more than 72% of Filipino digital customers using social platforms to research and discover brands, with up to 86% open to strong creator influence. Those are conference presentation figures with the underlying study unnamed, so hold them as directional. The slide line: direct traffic is an artefact of the exposure layer, and it needs a named companion metric.

Put Exposure, Behaviour and Outcome on One Page

Let's take the replacement in order. The Search Engine Land analysis sets out a triangulation rule worth adopting wholesale: combine three signals with different blind spots, an exposure metric, a behavioural signal and a business outcome, report all three on one client reporting page every month, and treat convergence as confidence and divergence as the thing to investigate. Three bearings, like a surveyor's, resolve a position no single one could give. Amanda Natividad, VP Marketing at SparkToro, advises planning for "the signals that happen before, or instead of, the visit", with measurement expanding to visibility, citations and mentions alongside platform signals and eventually conversions.

Exposure. Mentions, citations and share of conversation, counted monthly, which is most of how monitoring work shifts from clippings to citations. Platform reach settles where to count. DataReportal's Digital 2026 Philippines report gives advertising reach as a share of the internet user base at 97.7% for Facebook, 67.1% for Messenger, 65.3% for TikTok and 60.8% for YouTube. Those are reach estimates rather than active-user counts, published with vendors Meltwater and We Are Social. A baseline weighted toward X or LinkedIn describes almost nobody here, while national, provincial and vernacular coverage describes the real conversation. Messenger and Viber volume stays largely unobservable, which belongs on the slide as a stated limitation. Keep sentiment directional: automated reads across Filipino and Taglish are improving, not settled.

Behaviour. Branded search, unreferred arrivals, engagement, and a CRM field capturing how a lead says they found you. These are the only instruments that register influence arriving without a referrer, and competitive reads from social data belong here too.

Outcome, sourced outside the platforms you report on. The 2026 Philippine Hyperlocal Study from Ogilvy, WPP Media and Fifty5Blue describes an SEC D and E mass market of 28.7 million people who discover inside apps, "rarely browse the open web", and among whom 100% of respondents rely on cash. No platform dashboard attributes a cash purchase to a click, so this layer of client reporting comes from sell-out or distributor data the client owns.

Platforms are filling the same gap with their own numbers. Google Ads has added video impressions to conversion paths and a "Performance by stage" card grouping its metrics into awareness, consideration and action. It excludes conversions imported from Google Analytics 4, and the source warns that an impression in a conversion path did not necessarily cause the conversion. That exclusion matters more to agencies than the feature does, because it drops exactly the clients whose outcomes live in their own analytics, which is most retainer clients here.

Incrementality Testing Now Has a Price Floor

The obvious objection is that incrementality testing already solves this, and it just got easier to reach. Conversion Lift measurement is now self-service for Search and Performance Max, spotted by Hana Kobzová of PPC News Feed and reported by Search Engine Roundtable. The entry conditions are the problem: at least 1,000 observed conversions, a US$5,000 budget, and one study at a time across Brand, Search and Conversion Lift.

Set that floor against local spend. HubSpot and Milieu Insight's Digital Marketing in Asia research found only about 20% of Philippine businesses spending over US$5,000 a month on all their digital marketing, from more than 1,000 SMB respondents across Singapore, Hong Kong, the Philippines and India in 2024. Treat it as indicative rather than a national statistic: single-sourced, no per-country respondent count published, and the publisher sells marketing software. Even read loosely, one lift study costs roughly four in five surveyed businesses more than a month of all their other digital work. The conversion threshold is the harder gate, and no local figure exists for it. My own inference, flagged as reasoning: a 1,000-conversion observation window assumes a monthly volume most Philippine retainer clients do not generate.

So incrementality is not the escape route for an agency running client reporting across a dozen accounts. The instruments that survive will be the cheap, manual ones, and the local reason sits in Swarm's Philippine AI Report with the Analytics and AI Association of the Philippines, covering 175 organizations: 65% of local AI initiatives never get past the pilot stage. A reporting layer that needs a new stack joins that graveyard. A monthly mention and citation count, owned by a named analyst, does not.

What Client Reporting Should Show at the Next Review

Demote the traffic slide. Do not delete it, because a client who has seen it quarterly for three years notices its absence before its decline. Put the exposure baseline underneath it this quarter, labelled as a baseline rather than a result.

Then resist importing the publisher playbook wholesale. Publishers revisiting registration walls are answering the same loss of visibility from the revenue side. Digiday reported The Verge rolling a hard registration wall to a cohort of readers, with audience and subscriptions director Esther Cohen saying "even if there's a loss of volume, there's hopefully a growth of quality", and one publishing exec putting ad rates for registered users at "as much as eight times".

The Philippine demand side does not support that logic yet. The Reuters Institute's Digital News Report 2026 chapter on the Philippines, by Yvonne T. Chua of the University of the Philippines, reports 13% of Filipinos paying for online news, down two percentage points year on year, with social media still the dominant source of news. Local reader-revenue work is real but earlier-stage: The Manila Times, ABS-CBN and the Philippine Daily Inquirer joined the first Southeast Asian edition of the Google News Initiative Digital Revenue Launchpad in 2023, and The Manila Times now runs a registered-free tier alongside paid ad-free tiers. Known-user trading is an open question here, not a trend, so do not plan a client's earned media around a growing pool of logged-in Filipino readers.

What the page carries at the next review: one exposure number with its surface named, one behavioural signal that catches unreferred arrivals, one outcome the client owns the data for, and a note on what each cannot see. That last line separates client reporting from a dashboard screenshot. If you want to see the shape of an exposure baseline before building one for a client, a sector brief from the report library shows what a month of a sector's conversation looks like counted this way.

Key takeaways

• Keep the traffic slide and move it down the deck. Alone, it describes a shrinking share of the journey.
• Open the exposure baseline this quarter. Mentions and citations cannot be backdated, so a rough baseline now beats a better method in April.
• Source the outcome layer outside the platforms you report on. For a mass-market client paying in cash, that means sell-out or distributor data.
• Weight exposure to Facebook, Messenger, TikTok and YouTube, and name the surfaces you cannot observe.
• Treat Conversion Lift as a large-client instrument: its US$5,000 floor alone exceeds a month of total digital spend for four in five surveyed Philippine businesses.