HubSpot tried to build the biggest data co-op in B2B software, then killed it four days later. On July 1, 2026, HubSpot shipped Contact Discovery: a feature that pools business-card data and email engagement signals from every customer's CRM into one shared commercial dataset. Sales teams evaluating intent data tools have watched this exact tension play out before — more shared signal makes every tool better, until customers realize they don't actually know what they agreed to share. By July 5, HubSpot had reversed course entirely. Adam Robinson, founder of RB2B, thinks that reversal was a historic mistake. We think the truth is more specific than either side wants it to be.
The HubSpot Data Co-Op That Lasted Four Days
Contact Discovery was pitched as a sales productivity feature: help reps find and verify new contacts faster by drawing on a shared commercial dataset built across HubSpot's entire customer base. Mechanically, it combined business-card-level data — name, title, company, email — with email engagement signals that show deliverability status, pooled from customer CRMs and anonymized. Strip away the productivity framing and what's left is a straightforward data co-op: every customer's CRM activity makes the shared pool better, and every customer benefits from everyone else's activity feeding it. That's not a new idea. It's the same logic behind every intent data platform that scores buying signal by aggregating behavior across accounts nobody individually owns.
The problem showed up almost immediately, and it wasn't about the idea. A community member posting as lexnicole asked six specific questions about how two settings interacted: 'AI Model Training' and 'enrichment.' The question was simple and completely reasonable — if a customer disabled AI Model Training, did that alone keep their data out of the shared enrichment pool, or did enrichment need to be turned off as a separate, second step? Nobody at HubSpot could give a clean answer fast enough. HubSpot's own Chief Product Officer later acknowledged that the original communication 'did not meet the standard you expect from us when it comes to transparency.' That's not a company defending its rollout. That's a company admitting its opt-out mechanism was genuinely unclear.
Four days after the terms took effect, HubSpot's Duncan Lennox published a full retraction titled 'We Got This Wrong. And We Are Fixing It': 'We made a mistake...we will not move forward with the terms of service changes we communicated on July 1, 2026.' HubSpot committed to redesigning the enrichment opt-in controls from scratch, with no new timeline given for Contact Discovery, which had originally been scheduled to launch broadly on August 4, 2026. Reported in detail by ppc.land, the whole arc — announcement to full reversal — took less than a week.
Adam Robinson's Verdict: HubSpot Folded to a Two-Week Tantrum
Adam Robinson, founder of RB2B and Retention.com, posted his reaction on X on July 23, 2026, and he did not hedge. Robinson has spent 2026 arguing that standalone signal and intent-data tooling is consolidating — his July 15 piece, 'The Death of Signal Orchestration,' named Koala, Warmly, Common Room, Pocus, and Unify as categories getting squeezed by warm-outbound tools that just build the signal in. His HubSpot take is the same instinct applied to a bigger dataset: whoever aggregates the most signal wins, and customer objections are noise that fades once the news cycle moves on.
“I think HubSpot made a HUGE mistake. They never should have backpedalled on their decision to create the greatest data co-op in the world. Here is the simple reality. People would have whined about it on LinkedIn for two weeks, a handful of customers would have left, then everybody would have moved on to the next thing, and HubSpot would have been left with the single greatest data co-op that has ever been created. This would have been INCREDIBLE for every HubSpot user.”
Robinson's framing rests on a specific bet: that customer objections to Contact Discovery were volume, not substance — LinkedIn noise that a company with HubSpot's market position could simply outlast. It's worth taking that argument seriously, because Robinson's read on where B2B data tooling is headed, covered in our breakdown of the signal orchestration debate, has been directionally right about consolidation elsewhere. The question isn't whether pooled data has value. It obviously does — that's the entire premise behind every intent-data platform sold today. The question is whether what happened here was actually a loud minority whining, or something more specific.
Something Inc.'s Verdict: The Backlash Was Right, the Upside Was Real
Read the actual complaint again. lexnicole didn't object to HubSpot building a shared dataset. She asked whether two named settings — AI Model Training and enrichment — controlled the same thing or two different things, and whether turning off one protected her CRM data or left a second door open. That's not outrage. That's a customer doing exactly what she should do before her company's CRM data feeds a system she doesn't control: read the settings, find them ambiguous, and ask HubSpot to clarify before assuming consent. HubSpot's own CPO didn't dismiss that question as noise. He conceded the communication failed on transparency. When the company holding the data agrees the opt-out was unclear, that's not a PR reflex — that's the company confirming the complaint was correct.
| DATE | EVENT | WHAT IT ACTUALLY SHOWED |
|---|---|---|
| Jul 1, 2026 | Contact Discovery terms take effect | Enrichment and AI Model Training settings ship without a clear, separate opt-out for each |
| Jul 1–4, 2026 | lexnicole posts six specific questions publicly | A customer identifies a real consent gap, not a vague complaint |
| Jul 5, 2026 | HubSpot publishes full retraction, CPO admits transparency failure | The company confirms the ambiguity was real, not manufactured outrage |
| Aug 4, 2026 (scrapped) | Original planned broad launch date for Contact Discovery | No new date has been announced |
That's where Robinson's argument breaks down and where it holds up, in that order. It breaks down on the premise that this was a PR problem HubSpot could have out-waited. You don't out-wait a consent design flaw your own CPO has publicly conceded. B2B data sharing built on a CRM — the system of record for a customer's actual pipeline, actual contacts, actual deal data — cannot ship on an ambiguous default and rely on customers not reading the terms closely enough to catch it. That's not resilience. That's exposure. The moment the ambiguity became public and HubSpot's own leadership validated it, the reversal wasn't optional.
Where Robinson holds up is on the underlying strategic bet. A genuine data co-op — pooled signal across a customer base large enough to make every individual member's tooling smarter — is a real advantage, and it's the same logic behind every serious intent-data platform on the market. HubSpot didn't lose because the idea of a data co-op was bad. It lost because it tried to build one with a consent model that even its own product leadership couldn't explain cleanly when a customer asked a direct question. Those are two separate failures, and only one of them is fatal. The idea survives. The execution didn't.
The New Standard for Intent Data Tools: Opt-In or Nothing
This matters beyond HubSpot because every vendor selling intent data tools is making the same bet HubSpot made, just with less public scrutiny. Signal orchestration platforms, enrichment vendors, and CRM add-ons all depend on some version of B2B data sharing across their customer base to make their scoring, their contact matches, or their intent signals better than what any single company could build alone. That's not a hidden business model — it's the whole value proposition. The HubSpot episode is useful precisely because it made the mechanism visible and testable in public, in a way most vendors' terms of service never get.
This is directly relevant to anyone running cold email programs that lean on enrichment or intent data to prioritize outbound targets. If the vendor supplying your signal is quietly pooling your CRM activity into a shared dataset with an unclear opt-out, that's not just a privacy question — it's a data-quality question, because a vendor that can't explain its own consent model probably can't explain its data provenance either. The same discipline that should make you skeptical of inflated AI outbound tooling ROI claims applies here: ask the vendor to show its work, not just its pitch.
Where This Leaves Intent Data Tools Buyers
Neither extreme survives contact with what actually happened. Robinson's version — that this was two weeks of LinkedIn whining HubSpot should have ridden out — ignores that HubSpot's own CPO conceded the complaint was accurate. But the instinct underneath Robinson's argument isn't wrong: a real data co-op, built with legible consent from day one, is a durable advantage in a market where signal orchestration tools are consolidating around whoever aggregates the most usable data. HubSpot's mistake wasn't ambition. It was shipping the ambition with a consent design its own team couldn't explain when asked a direct question.
Do this next: before you renew or sign with any intent data or enrichment vendor this quarter, ask them in writing whether your CRM or contact data feeds any dataset shared across their customer base, and get the opt-out mechanism described in one paragraph you could hand to a customer without embarrassment. If they can't produce that paragraph cleanly, treat it the same way HubSpot's own customers correctly treated Contact Discovery's terms — as a gap, not a formality. Pair that audit with a hard look at whether the tool is actually driving pipeline or just adding a line item; our take on whether cold email is still worth it in 2026 and our comparison of AI dialers against cold email budget both start from the same premise: judge outbound infrastructure on what it proves, not what it promises. If you sell into tech and SaaS accounts, your buyers are reading the same HubSpot story you are. They will ask you the same six questions lexnicole asked HubSpot. Have the answer ready before they do.
See where you are cited today
A free snapshot audit of your rankings and AI citations before we ever talk.
Josh leads work at the intersection of SEO and generative engines at Something Inc., helping B2B brands get ranked and cited across every major AI engine.