Blogs

What is feed exposure intelligence? The missing layer in consumer research

Research Methods
7 Minute Read
Feed exposure intelligence measures what personalized feeds actually show people, item by item, linked to who they are. What it is, how it differs from social listening and media measurement, and what it can't tell you yet.

What is feed exposure intelligence?

The missing layer in consumer research

Written by: Mark Nguyen, ND Research

Originally Published: October 23, 2026

Labels: Research Methods, AI-assisted, Continuously Maintained

Topics: Feed exposure intelligence, Consumer research, Social listening, Media measurement, Recommender systems

TLDR: Most consumer research measures what people say, buy, click and post. Almost none of it measures what the algorithm put in front of them first. Feed exposure intelligence is our name for measuring that.

Key Ideas:

  1. Most consumer research measures the trail people leave behind: what they say, buy, click and post. Very little measures what the algorithm put in front of them first.
  2. Feed exposure intelligence is our name for measuring that input: the items a personalized feed actually served, linked to consented, profiled people.
  3. It sits between social listening (public posts, in aggregate) and media measurement (which apps and channels, for how long).
  4. Exposure is not attention, persuasion or purchase. We use a six-level ladder to keep those apart.
  5. It's early. Our method is patent pending and in private beta, and independent validation is still ahead of us.

‍

Preamble

Every few weeks someone asks us some version of the same question: is this social listening? Is it media measurement? Is it a panel? Fair questions, because the thing we're building doesn't have a category yet, and a product without a category is super hard to explain (we've learned this the hard way).

This is our attempt to define the term we use, feed exposure intelligence, carefully enough that you could disagree with it. Where we're still guessing, I've tried to say so.

‍

What is feed exposure intelligence?

Feed exposure intelligence is the measurement of what personalized social and recommendation feeds actually show people, item by item (creator clips, memes, news posts and ads alike), linked to who those people are. It's the name we use at ND Research for the category our product, Tapestry, belongs to.

The shortest version we have: everyone measures the ad, and we read the feed around it. (Think of it as a blood test for a feed, if you will.)

Think: two people open the same app at the same time. One gets gym content, debt-payoff creators and a car ad. The other gets K-beauty tutorials, comedy dog videos and the same car ad. Same app and same ad, but a very different internet. If you only measured the ad, those two people look identical.

‍

What does consumer research miss today?

Mostly the inputs.

The mainstream research toolkit is actually very good at outputs:

  • Surveys tell you what people say.
  • Purchase data tells you what they bought.
  • Clicks and searches tell you where attention had already landed.
  • Public social posts tell you what people chose to broadcast.

Each of these is the result of a decision. What the feed served beforehand is one of the inputs to it, and recommendation systems now pick a lot of that input for each person individually. Consumer researchers have argued for decades that identity is built partly through what we consume (Belk's 1988 "extended self" is the classic reference [1]).

Our working hypothesis is that a personal feed is a big part of that story now, and that repeated exposure quietly shapes preference. It's a hypothesis, and we treat it like one.

Media measurement has traditionally focused on usage (which app, which channel, how many hours) rather than the individual items inside a personalized feed. In our own conversations this year, a measurement leader at a global research firm told us their device meters can see which social apps someone opens but not what those apps show them.

And platform-level data, where it's available at all, is expensive. A senior executive at a large advertising holding company told us even his group couldn't afford it (more on those conversations in our notes from 19 industry conversations [2]).

‍

Why does this matter for the market research industry now?

Because the industry is growing slowly while more of the consumer's world moves inside feeds that are hard to see from outside the platforms.

‍

$36.4 billion - IBISWorld's estimate of US market research industry revenue in 2025, forecast to grow about 2.4% a year to 2030 [3]

‍

IBISWorld estimates US market research industry revenue at $36.4 billion in 2025, growing about 3.8% a year from 2020 to 2025, with a forecast of about 2.4% a year from 2025 to 2030 [3]. The same report points to advertising on social platforms like TikTok and Instagram as one of the things driving demand for research.

So demand is coming from a channel the traditional toolkit doesn't observe very well. That's the opening we see, and to be frank, it's why we started building this in the first place.

‍

How is feed exposure intelligence different from social listening and media measurement?

The simplest way to see it is the unit each tool observes:

  • Surveys and panels: answers from a person (what they say).
  • Social listening: public posts matched by keyword or topic, in aggregate (what people say in public).
  • Media and audience measurement: apps, channels and ads, often with time spent (where people were).
  • Ad and creative databases: which ads ran (what advertisers put out).
  • Feed exposure intelligence: the specific items a person's feed served, ads included, linked to that person's profile (what each person was shown).

These categories overlap, and vendors are moving toward each other. We see feed exposure as the layer in between, and it works best alongside them.

‍

How does Tapestry measure feed exposure?

Here's the method as it works in our private beta today:

  1. Consent. Adults (18 and over) opt in. Their uploads are saved under a study number, not a name.
  2. Capture. Participants share screenshots of their own feeds from two of six supported platforms (Instagram, TikTok, YouTube, Facebook, X and Reddit), at least five per platform. We're also testing other ways to capture that ask less of participants.
  3. Classification. Each item is labelled against the IAB Content Taxonomy (version 3.1, tiers 1 to 3) [4] and given one of five "uses and gratifications" labels: Identity Formation, Information Gathering, Aesthetic Experiences, Entertainment and Relaxation, or Social Connection. That last one is a read on why someone might be scrolling it, not just what it is. For survey studies, we're adding each item's format (e.g. short-form video, carousel, story) and whether it's an ad.
  4. Linking. Items connect to the participant's demographics and survey answers (for example, which products they use, when a study asks).
  5. Reporting. Strategists see aggregate reports: how widely a brand, creator or topic shows up in a target audience's feeds, whether two brands reach the same viewers, who over-indexes, how concentrated exposure is, and (in beta) how a planned ad fits the feed it will land in.

Reports are built on aggregate findings, and panel participants' screenshots are hidden from client views. Our viewer profile and distribution reports hide groups of fewer than 30 viewers, to make it harder for a cut to point back to a person.

‍

How do shopper metrics translate into feed metrics?

If you've worked with retail or household panel data, the shape will feel familiar. Here's the translation we use:

  • Household penetration becomes viewer penetration: the share of people whose feed contained a brand, creator or category.
  • Purchase frequency becomes exposure frequency: how often it shows up for the people who see it.
  • Units per trip becomes exposure per session.
  • Share of basket becomes share of feed.
  • Repeat purchase becomes repeat exposure.

Two of these, true exposure frequency and repeat exposure, need the same people across several sessions, which we're only starting to collect. The big difference from the shelf is that the shelf is personalized. Basically, every person walks into a different store.

‍

Does exposure mean someone paid attention?

No, and we try hard not to blur the two. We use a six-level ladder:

A six-step ladder from availability to cognitive consumption, with levels 3 (recognized exposure) and 4 (personified exposure) marked as measured today
Figure 1: The exposure ladder. Tapestry measures levels 3 and 4 today; attention and meaning are not measured. NDR, October 2026.

‍

  1. Availability: the item exists and could be served.
  2. Device delivery: it was sent to someone's device.
  3. Recognized exposure: a specific, identified item appeared on their screen.
  4. Personified exposure: that exposure is tied to a known, consented, profiled person.
  5. Attention: they actually looked at it.
  6. Cognitive consumption: it registered and meant something to them.

Today, Tapestry measures levels 3 and 4. We get at the "why" indirectly, through the gratification labels and what participants tell us, but we don't claim to measure attention or persuasion. Exposure doesn't equal endorsement, engagement or conversion, and our reports say so.

‍

Who is feed exposure data useful for?

The early interest we've seen comes from a few places:

  • Creative strategists and content teams who want to see what their work will sit next to: the creators, memes and competing ads in the actual feed.
  • Researchers and insights teams who want observed behaviour alongside attitudes, especially for segments that surveys struggle with (Gen Z is the one we hear about most).
  • Media and ad-tech people interested in the context around an ad. That came up in 6 of the 19 industry conversations we recorded this year, from agency, ad-tech and research voices [2].

Which of these turns out to be the main use is honestly still an open question for us.

‍

What are the limits?

Plenty, and we'd rather say them out loud:

  • It's early. Tapestry is in private beta. Samples are small, and results are unweighted.
  • Screenshots are a snapshot. A handful of screens per platform is a sample of a feed, not a complete record of it. How we capture is still evolving.
  • Classification is automated. A computer labels every item, and people check quality. It makes mistakes.
  • Validation is ahead of us. Independent tests of convergent validity, test-retest reliability and audited classification accuracy are the next big piece of work.
  • Exposure isn't cause. Seeing more of something doesn't prove it changed what someone bought.

‍

FAQ

Is feed exposure intelligence the same as social listening?

No. Social listening analyzes public posts by keyword or topic, in aggregate. Feed exposure intelligence looks at what consented people were actually shown in their own feeds, ads included, linked to who they are.

How is participant privacy handled?

Participants are adults who opt in. Their uploads are stored under a study number rather than a name, reports are built on aggregate findings, panel participants' screenshots are hidden from client views, and participants can ask to have their data deleted.

Which platforms does Tapestry cover?

Six: Instagram, TikTok, YouTube, Facebook, X and Reddit.

Is the method patented?

It's patent pending. The US application was published in March 2026 as US 2026/0087516 A1 [5].

How many people do you need for a feed exposure study?

Our Viewer Profile report hides groups under 30 viewers and flags groups under 50 as directional. For decisions you plan to act on, an advisor has pushed us toward roughly 150 to 300 people per segment.

‍

How to cite this

Nguyen, M. (2026, October 23). What is feed exposure intelligence? The missing layer in consumer research. ND Research. https://www.nguyendigital.co/blog-posts/what-is-feed-exposure-intelligence

‍

List of References

(Continuously Maintained)

[1] Journal of Consumer Research. Belk, R. W. (1988). Possessions and the extended self. 15(2), 139-168.

[2] ND Research. Nguyen, M. (2026, October 23). What does the insights industry actually struggle with? Notes from 19 recorded conversations in 2026. 19 recorded conversations with 17 people and organizations, April 1 to October 20, 2026: agency and holding-company leaders, a brand-side creative lead, research and measurement professionals, independent advisors and consultants, and social listening and synthetic-data vendors. The appendix of that article lists each conversation by month, role and context. https://www.nguyendigital.co/blog-posts/what-strategists-and-insights-teams-struggle-with

[3] IBISWorld. Govdysh, A. (2025, October). Market Research in the US, industry report 54191.

[4] IAB Tech Lab. Content Taxonomy, version 3.1. https://iabtechlab.com/standards/content-taxonomy/

[5] United States Patent and Trademark Office. Patent application publication US 2026/0087516 A1, published March 2026.

Still curious?

Incomplete and outdated insights are a thing of the past. We give you a front seat into how your customers live, work and interact. We weave together diverse inputs into a complete picture of your product.