Most of what gets called “B2B data” is a commodity. If your competitor can rent the same list, the data is not a moat. It is a line item.
I am George, founder of Leadpipe. I spend a lot of time thinking about the difference between data you can buy and data you can only build, because that distinction determines whether a go-to-market stack produces advantage or parity. The short version: the commodity data market is where most budget is spent and most returns are diluted. Non-commodity data is where the actual leverage is, and most teams do not have any.
Here is a working definition.
Commodity data is any record your competitors can rent on the same terms.
The test is simple. Can a direct competitor sign a contract with the same vendor and get functionally the same data? If yes, the data is a commodity.
This is most of the B2B data market. ZoomInfo contacts, Apollo lists, Cognism numbers, LeadIQ mobile data, Lusha extensions, Clearbit firmographics (now Breeze Intelligence inside HubSpot), RocketReach emails. All commodity. The coverage varies, the prices vary, the guarantees vary, but the underlying proposition is the same: here is a database, here is the contract, pay us and draw from the same well as everyone else.
| Commodity data source | What you get | What your competitor also gets |
|---|---|---|
| ZoomInfo | ~321M contacts, firmographics | Same 321M contacts |
| Apollo | ~275M contacts, email accuracy ~95% (claimed) | Same 275M contacts |
| Clearbit / Breeze | Firmographic enrichment via HubSpot | Same enrichment |
| Cognism | EU/UK-strong GDPR data | Same EU/UK data |
| LeadIQ | ~600M contacts, LinkedIn-driven | Same 600M contacts |
There is nothing wrong with commodity data. It has legitimate uses (enrichment, TAM, account planning). But the go-to-market team that depends on commodity data for differentiation is running a race where everyone has the same car.
Commodity inputs produce commodity outcomes. This is why we bought ZoomInfo and did not see ROI: not because the database was bad, but because every competitor was running the same play on the same contacts.
Non-commodity data has one of three sources: first-party, live behavior, or proprietary resolution.
The data that produces actual advantage comes from three places. They are not mutually exclusive, and the best stacks combine them.
Source 1: First-party data from your own infrastructure.
Every interaction with your website, app, docs, support, and community produces first-party data. Most teams capture 10-20% of the signal available and waste the rest. GA4 tells you page views but not who visited. Your form captures 2-3%. The other 97-98% of the visit data is discarded.
Building a first-party layer means:
- Visitor identification on every marketing page, not just the form-gated ones.
- Session-level behavior captured: pages, duration, return pattern, session chains.
- Attribution tied to actual identity, not cookies.
- Product telemetry routed into the same identity graph.
Your traffic is not someone else’s traffic. The intent behavior of your visitors is not available to your competitors. This is the most valuable dataset most B2B companies sit on top of and systematically ignore.
Source 2: Live behavioral networks beyond your own properties.
Your first-party data is rich but bounded to your own surface. Live behavioral networks extend the surface to include cross-site research behavior. Person-level intent data via Orbit works by observing behavior across 5M websites and resolving it to identified individuals.
This is different from traditional intent data providers (Bombora, G2 Buyer Intent). Those typically work at the company level (account-level intent) and refresh on weekly cycles. Person-level intent works at the individual level, daily refresh, across topics beyond just review-site visits.
A team that knows “Sarah Chen at Acme Corp read three articles on data privacy this week and visited two vendor sites” has non-commodity signal. A team that knows “Acme Corp showed medium intent in the data privacy category this month” has commodity intent. The person-level version is actionable. The account-level version is directional.
Source 3: Proprietary identity resolution.
The third source is the resolution layer itself: the ability to turn a browser session, a hashed email, a mobile ad ID, or a first-party signal into a verified identity. This is what Leadpipe’s identity graph is.
We built it because we believe licensed graphs are a dead end. Anyone can license the same graph. The graph becomes a commodity. A graph you build from first-principles, with your own collection methods, your own consent framework, and your own verification pipeline, is not renewable through a procurement process. It has to be built.
The math on graph quality is stark. Our independent accuracy test scored deterministic identification at 8.7/10, against probabilistic identification at 5.2 (RB2B) and 4.0 (Warmly). The same traffic, different resolution methods, dramatically different accuracy. The difference is whether the graph is built or inferred. Commodity resolution layers tend to be inferred. Non-commodity ones are built.
Non-commodity data has five practical properties.
Abstract definitions are less useful than a checklist. Here is how to tell whether a given data source is commodity or non-commodity in practice.
- Source-unique. Is the underlying signal observable only from this source, or is it available from multiple vendors? Non-commodity data has a single observable source.
- Fresh. Refreshed on a cadence (hourly, daily) that reflects the reality of the underlying behavior, not the vendor’s convenience (quarterly dumps).
- Behavioral. Includes what the person or company is actually doing right now, not just what they are. Fit is not intent.
- Actionable at person level. Resolves to a specific individual with enough context (pages, topics, timeline) to support a contextual message, not a generic reach-out.
- Contract-defended. The data provider has terms that prevent competitors from accessing the same customer-facing insights. This is not about secrecy. It is about whether a competitor gets the same output if they sign.
A data source checking 4-5 of these is non-commodity. A source checking 0-1 is commodity. Most of the B2B data stack as deployed today is in the 0-2 range.
The economics look different for non-commodity data.
| Dimension | Commodity data (ZoomInfo, Apollo, Clearbit) | Non-commodity data (Leadpipe, Orbit, first-party) |
|---|---|---|
| Price per record | $0.01-0.20 | $0.12-0.30 |
| Freshness at use | Days to months | Hours |
| Signal density per record | Demographic only | Behavioral + demographic |
| Reply rate on outreach | 1-3% | 15-25% |
| Cost per meeting | $300-600 | $15-60 |
| Availability to competitors | Full | Bounded to your own traffic and network |
The per-record price of non-commodity data is slightly higher. The cost per meeting is an order of magnitude lower, because the signal density per record is different. You do not need to send 500 messages to book a meeting. You send 5-20 to people who are actually in-market.
This is the inversion that changes the unit economics of outbound. The commodity stack optimizes for contacts per dollar. The non-commodity stack optimizes for conversations per dollar. These are different metrics with different winners.
The steelman: “Commodity data is good enough if the execution is great.”
Strongest counter: “Execution matters more than data source. A great SDR with a mediocre list will outperform a mediocre SDR with a great list. Focus on the craft, not the input.”
Partial credit. Execution matters. Great operators outperform on any input. But two things.
First, great operators produce compounding returns on non-commodity data that they cannot produce on commodity data. A great SDR running an identified-visitor workflow books 3-5x the meetings of the same SDR running ZoomInfo cold lists. The ceiling is different, not just the floor.
Second, the market is efficient enough that the execution gap between teams narrows over time. Everyone hires from the same pool. Everyone runs the same playbooks (because those playbooks are also commodity). The only durable source of advantage is the data layer. Teams that ignore this end up with great execution on a commodity input, which is parity, not advantage.
The exceptions are teams with proprietary channels (founder-led networks, exceptional community-driven funnels, unique distribution positions). For those teams, execution genuinely dominates. For everyone else, the data layer is the limiting factor.
Commodity inputs produce commodity outcomes. If your data is available to your competitor, your advantage is not in the data.
What this means for your week.
Four moves.
- Inventory your data stack by commodity score. List every data source feeding your outbound, marketing, and sales stack. For each, apply the five-property checklist above. How many sources check 3+ properties? For most teams, zero.
- Identify one non-commodity signal you are not using. Usually it is your own traffic. Install visitor identification. Your competitors do not have access to your traffic. That is definitional.
- Rewire one workflow from commodity to non-commodity. Take the worst-performing cold sequence and replace the input list with identified website visitors plus behavioral triggers. Measure reply rate and meetings over 30 days.
- Reduce your commodity spend by 20%. Most teams are overpaying for commodity data because nobody has audited the spend against outcomes. Cut the seats that are not producing, redirect budget to signal sources.
The trade is not dramatic in year one. Over 18-24 months, it compounds into a go-to-market advantage competitors cannot rent their way past.
The bottom line.
Most B2B data is a commodity. That is not a moral problem. It is a strategic fact. The teams building non-commodity signal layers (first-party behavior, live intent networks, proprietary identity resolution) are building advantages that are not available through procurement.
We built Leadpipe and Orbit to be the non-commodity layer for go-to-market. Not a feature on a platform. Infrastructure. The difference shows up in the pipeline, the deliverability, the response rate, and the close rate. It is not subtle.
If you want the short version: $147/mo gets you person-level identification on 500 visitors with full contact data. See full pricing →
