New: tag your media, then ask your data anything — in Claude, ChatGPT or Gemini. Start free with $500 in credit →
Analytics for the LLM era · real-world outcomes included

Tag your media. Ask your data anything.

WebsiteCTV & streamingProgrammaticOut-of-home

Quorum turns your web, CTV, programmatic and OOH logs into one structured dataset an LLM can actually use — resolved to ~74.6M real U.S. households, and carrying the dimension analytics never had: whether they walked in. Drop a pixel, get your data back as an asset you own, and ask it questions in plain English.

74.6M
Households
736,984
Venues tracked
100%
Deterministic
Opt-in
· k-anonymous
Inc. 5000 — America’s Fastest-Growing Private Companies
Honoree 2024 · 2025 · 2026

How it works

Three things that have never been joined. One graph that joins them.

Media exposure, website traffic and real-world visits — resolved to the same household. Every pathway between them is a question you can finally answer. And the whole graph sits inside your environment: queried by your own LLM over MCP, and matched against your own first-party data under clean-room controls.

YOUR LLM queried over MCP YOUR 1P DATA cleanroom-matched ONE HOUSEHOLD MEDIA EXPOSURE What you paid to show CTV & streaming Programmatic & display Out-of-home WEBSITE TRAFFIC Who landed, and from where Google Bing YouTube CTV & programmatic Direct & OOH-driven REAL-WORLD VISITS Who actually showed up Your locations Competitors Every major B&M brand Cross-shop & conquest 736,984 tracked venues every pathway is one query, not one vendor — and it runs where your data already lives

Plan

Before you spend: which publishers, day-parts and markets over-index on the people who actually visit your category.

Report

What happened, in outcomes rather than proxies — web visits and store visits per impression, per dollar.

Optimize

Mid-flight, not post-mortem. Shift weight to the inventory converting in the real world while the flight is still live.

Target

Turn the finding into an audience — real visitors, suppressions and lookalikes, pushed to your seat.

The outcome

See what is working and what is not — against real-world action, not clicks, views or modeled proxies. The same number for every channel, so the comparison is finally fair.

And then

Mine the deeper pockets. The publishers, creatives, day-parts, zip codes and store clusters that over-index quietly — the opportunity nobody found because nobody could see this join before.


One platform, every pair

Every question you have is an exposure crossed with an outcome.

Today each of these is a different vendor, a different study, a different six-week wait. On one household spine they are the same query with different arguments — and they compose.

→ Website visit → Store visit → Audience to activate
CTV & streaming Which shows and apps send people to my site — and at what rate versus the households that never saw the ad? Which CTV inventory drives walk-ins. Publisher, creative, day-part and store, mid-flight instead of post-mortem. Retarget the households that saw the spot but never showed up — or suppress the ones who already did.
Web & programmatic Is this traffic real? How much of what you paid for resolves to a real, in-market U.S. household. Which on-site behavior predicts a visit — which pages, which sources, which segments actually convert offline. Build a lookalike from your visitors who also visit physical places, not from a modeled interest bucket.
Out-of-home Did the board move search and site traffic? Exposure measured from the viewshed, not a tag. Did the board move footfall — to your door, and away from the competitor down the block? Reach the commuters a placement actually exposed, on CTV and the open web.
Your 1P data Match your CRM under clean-room controls and see which customers came back. Where else your customers shop — cross-shop, conquest, and which competitors are taking share. Grow it: audiences that look like your best real-world customers, pushed to your seat.

Same pixel. Same household graph. Same plain-English interface. No new integration per question.


Get started

Three steps. Your first million impressions are on us.

No sales call to see it work. The public tools answer questions right now with no account at all; the account is what turns your own media into a dataset.

01
🔌

Try it with no account

Add mcp.quorum.live/mcp as a connector in Claude, ChatGPT or Gemini and ask about any brand’s shoppers, foot traffic or cross-shop. Five tools, no login, no card.

02
🏷️

Generate your pixels

Create an account and we issue your web pixel and your impression pixel — one tag for your site, one tracker URL for any DSP, with the macros pre-filled per platform. Paste and go.

03
💬

Ask your own data

Within a day of first traffic your data is household-resolved and queryable — in the app, over the REST API, or by asking your LLM. Store visits attach as soon as your locations are matched.

<script src="https://d27vf9ijvhczsn.cloudfront.net/q.js" data-adv="YOUR_ID"></script> https://d27vf9ijvhczsn.cloudfront.net/pixel.gif?adv=YOUR_ID&imp=${AUCTION_ID}

Your real tags come pre-filled with the right macros for The Trade Desk, DV360, Xandr, MNTN, Roku, Yahoo and FreeWheel.

$500 in credit — one million impressions

We measure at $0.50 per thousand impressions, so $500 is exactly 1,000,000 impressions household-resolved and measured — no rounding, no asterisk — plus unlimited questions against them. No card up front. Keep going on a plan, or stop and keep the report.


The LLM-native interface

Google Analytics for the LLM era — plus whether they walked in.

Not another dashboard to learn. Quorum is a structured dataset your model can query — the same web traffic an analytics tool sees, plus the channels it never saw and the outcome it never had. Ask in plain English, or wire it into your own LLM views and join it against your 1P data under clean-room controls. The panel here is not a mockup: it is querying our live MCP server right now.

  • "Of the households our CTV reached, how many converted in the real world?"
  • "Match this log file to households and append physical-behavior features."
  • "Score this audience's real-world quality before we activate it."

MCP connector: mcp.quorum.live/mcp · aggregate · privacy-safe · read-only

Live now · from mcp.quorum.live
🏆 The real-world conquest map — where a brand's shoppers over-index vs. the average consumer (not just who's biggest).
Where do shoppers over-index?
…or drop in any brand:
Asking Quorum…
🌐 Explore the full Brand Universe →

Audience builder

Or start from the other end: buy the audience.

You do not have to tag anything to get value on day one. Pick the physical retailers, restaurants and venues whose real visitors you want to reach — from the 736,984 venues Quorum already tracks — and we deliver that audience to your seat, with a report explaining exactly who is in it.

  1. 01Select the places. Any brand, category or venue set we already track — Home Depot visitors, Chick-fil-A regulars, Audi dealership shoppers, ski resorts in Colorado.
  2. 02Scope it. Recency window, national or DMA, state, or a radius around your own stores. Reach estimate updates as you go.
  3. 03Push it to your DSP. The Trade Desk directly, or LiveRamp fan-out to Meta, DV360, Xandr and the rest. Your seat, your line items.
  4. 04Get the description report. Emailed with every build: who is in the audience, where they live, what else they visit, and how it compares to the average consumer — so you can defend the buy.
Audience preview
Seed venuesHome Depot · Lowe’s
Visited within90 days
GeographyDallas–Ft. Worth DMA
Also over-indexesTractor Supply 3.4×
DestinationThe Trade Desk
Est. households412K
Illustrative. Live reach estimates come from the builder; every audience is k-anonymous with a floor of 50 households.
It compounds with your pixels. Audiences built from someone else’s footfall are good targeting. Audiences built from footfall plus your own web and impression logs are a closed loop: suppress who already converted, retarget who saw but did not show, and rebuild the seed weekly from who actually walked in. That is the same platform — you just turn the pixels on.

Pricing

One unit. Resolutions.

A resolution is one event — an ad impression or a pixel hit — matched to a real household, priced at a flat $0.50 CPM. That is the work, so that is the meter. Questions are unlimited: asking your data costs us almost nothing, so we do not bill it and we do not want you counting.

Every plan does everything

There is no feature you unlock by paying more. Tag every channel, measure every outcome, build and push audiences, match your 1P data — on day one, on every plan. The only difference is how much you can run through it.

Web pixelCTV & impression pixelsOOH exposure Store-visit attributionTraffic gradingCross-shop & conquest Audience buildDSP + LiveRamp delivery1P match, cleanroom-controlled MCP + REST + unlimited queries

Free

$0 no card
Prove it on your own data before you pay anything.
  • 100K resolutions/mo
  • Every capability above
  • 1 brand, 1 seat
  • Community support
Start free
$500 CREDIT

Launch

$299/mo
One brand proving one channel.
  • 700K resolutions/mo
  • Every capability above
  • 1 brand, 3 seats
  • Email support, 1 business day
Start on credit →

Growth

$1,500/mo
Running CTV, OOH and audiences together.
  • 4M resolutions/mo
  • Every capability above
  • Multi-brand, 10 seats
  • Shared Slack channel
Talk to us

Scale

$6,000/mo +
Agencies, networks and platforms reselling it.
  • 20M+ resolutions/mo
  • Every capability above
  • Snowflake share, private MCP endpoint
  • SSO, SLA, named contact
Talk to us

Scale’s extras are how you are served — a Snowflake share, a private endpoint, SSO and an SLA — not what the platform can do.

What a plan actually covers
Your $500 in credit
= 1,000,000 impressions
exactly, at $0.50 CPM
A site doing 250K visits/mo
≈ 250K resolutions
inside Launch, a third of it
An always-on national campaign
≈ 20M resolutions/mo
Scale, $6,000/mo

List price is $0.50 per thousand ($500 per million), and plans include volume below it — effectively $0.43 per thousand on Launch, $0.38 on Growth, $0.30 on Scale. Beyond your plan you pay list, dropping to $0.40 above 4M and $0.30 above 20M. Nothing auto-charges — you are emailed at 80% and the meter stops at 100% until you raise it.
A flat CPM on measured impressions is deliberate: you already buy media in CPMs, so you can put our line item next to your media line item and compare them without a conversion. One thousand impressions measured, half a cent each.

Audience Offset

Let your audience spend pay your measurement bill.

When you buy Quorum audiences in The Trade Desk, the data fee shows up in your own platform reporting — so you can already see exactly what you spent with us. We credit a share of that against your Quorum invoice, up to the whole thing.

Buy the audiences, and the measurement that makes them smarter starts paying for itself. Verified against your TTD reporting, credited monthly, and capped at your invoice — never a rebate.

Ask about Audience Offset

Available on any paid plan with a Trade Desk seat. Credit rate is set in your agreement and computed from The Trade Desk’s own usage reporting.

Add-ons: audience delivery $500 per million records pushed to your seat (about $50 per 100K). Custom venue mapping for locations we do not yet track: $1,500 per brand build, one time — that one is priced separately because our cartography team draws it by hand.


Real-World Movement Index · proof, not promise

Who America actually showed up for, year over year.

Year-over-year change in each brand's share of its category's store visits — deterministic foot traffic across ~74.6M opt-in U.S. households (H2 2024 → H2 2025). Free, public, and re-runnable in diligence — pick a category, or see how the Cannes Lions 2026 entrants moved.

A real-world result reflects both the creative and the media weight behind it — a brand can grow visits with a better idea, or simply by buying more exposure. Quorum measures the deterministic outcome (who actually showed up); paired with media-spend data, it isolates true creative efficiency — real-world action per dollar.

Browse the full Movement Index → · 82 brands, one page each, free to read and cite

Method: share = each chain's portion of total category store visits, resolved deterministically from opt-in device visits. Change is same-store (only locations active in both periods) and outlier-trimmed (implausible swings excluded) so it reflects real visitation, not coverage or sample changes; category-relative, controlling for seasonality and panel growth. Aggregate, k-anonymous — never PII. 🏆 = entered in Cannes Lions 2026; movement is not attributed to creative.

Independent by design

One real-world number everyone can trust.

Built entirely from 100% opt-in device signals, resolved to ~74.6M households, reported as k-anonymous aggregates — independent of every platform it measures, auditable, never PII. We refine your data; we never take it. Honest coverage on every answer, never "we see everyone." That's how it earns a place in your stack.

Plugs into the stack you run

Infrastructure, not another dashboard.

Quorum enriches the platforms and partners already moving your spend — refining their data instead of replacing it. Built to plug into the modern programmatic & CTV stack.

MagniteMNTNAttainCausalHearstPublicisParamountLotlinx& more
Native & interoperable

Built on Snowflake, AWS & Anthropic, and wired into The Trade Desk & LiveRamp — so Quorum meets your data where it already lives and activates where you already buy. Share securely through Snowflake, deploy in your own cloud, and query in plain language over MCP.

SnowflakeAWSAnthropicThe Trade DeskLiveRamp

Questions

The things people actually ask.

How do I measure whether my CTV ads drove store visits?

Point your DSP impression tracker at a Quorum pixel URL, or send us your impression logs. We resolve the exposed devices to households in a ~74.6M-household opt-in graph, then check those same households against deterministic visits to your locations. You get a visit rate for exposed households versus an unexposed baseline, broken out by publisher, creative, day-part and store — readable in the app, over the API, or by asking your LLM.

Can Claude, ChatGPT or Gemini query my marketing data?

Yes. Quorum runs an MCP server at mcp.quorum.live/mcp. Five core tools work with no account and no authentication, so any MCP-capable agent can start asking immediately. Adding your Quorum token unlocks your own tagged data: attribution, pixel health, traffic grading and audience building.

What is the “LLM version of Google Analytics”?

A structured, agent-queryable dataset instead of a dashboard. We ingest website pixel traffic the way an analytics tool does, but resolve it to households, add the channels analytics never saw (CTV, programmatic, out-of-home) and add the outcome analytics never had (physical store visits). You ask questions instead of clicking reports, and every answer is aggregate and k-anonymous.

How do you measure out-of-home and billboard exposure?

You cannot pixel a billboard, so OOH exposure is measured from device presence in the placement’s viewshed rather than from a tag. Those exposed devices resolve to households and then run through exactly the same outcome measurement as any other channel — website visits and store visits.

Can I build targeting audiences from store visitation?

Yes — it is a product on its own. We track 736,984 hand-drawn venues, so you can select the retailers, restaurants and venues whose visitors you want, scope by geography and recency, and deliver to The Trade Desk directly or through LiveRamp to most other platforms. Every audience ships with a description report: who is in it, where they live, what else they visit. Open the builder →

Do I need store locations to get value?

No. Web-only advertisers use Quorum for traffic quality, visitor profiling, cross-shop and audience building. If you do have locations and they are already among the venues we track, store-visit attribution turns on immediately; if not, our cartography team maps them (typically a one-time build).

How is this privacy-safe?

Every signal is 100% opt-in. Answers come back as k-anonymous aggregates with a floor of 50 households — no PII and no household-level records leave the platform. We refine the data you give us; we never resell it. First-party matching runs under clean-room controls, and independence is methodological: the objectivity does not depend on who owns us. Privacy policy →

How much does Quorum cost?

Every plan has every capability — the only difference is how much you can run through it. We measure at a flat $0.50 CPM ($500 per million impressions resolved), and new accounts start with $500 in credit — exactly one million impressions, no card. Plans include volume below list: Free $0 for 100K/mo, Launch $299/mo for 700K, Growth $1,500/mo for 4M, Scale from $6,000/mo for 20M plus a Snowflake share, private MCP endpoint, SSO and an SLA — which is how you are served, not what you can do. Questions are never metered. If you buy Quorum audiences in The Trade Desk, Audience Offset credits a share of that data spend against your invoice. Full pricing →


Reporting: app.quorum.live/login · API docs: api.quorum.live/documentation · MCP: mcp.quorum.live/mcp · Movement Index: quorum.inc/index · Audiences: quorum.inc/audiences · Sales: sales@quorum.inc · Privacy: app.quorum.live/privacy