Dreamdata

What it is

Dreamdata is a B2B revenue attribution platform that joins ad spend, website sessions, marketing automation, CRM records, and sales activity into one account-level customer journey, then attributes pipeline and closed revenue back across that journey. Copenhagen-based, founded 2018, and it sells itself as an agentic B2B attribution platform whose AI operates inside the attribution rules you define rather than inventing its own.

It is the second attribution vendor in the ooligo catalog and the warehouse-grade pole of the category. Where HockeyStack puts a fleet of GTM agents on top of the attribution graph, Dreamdata’s bet is that the graph itself is the product: a governed account-based model you can export, query with your own agents, and defend line by line to a CFO.

Dreamdata raised a $55M Series B led by PeakSpan Capital on 14 October 2025, with InReach Ventures, Angel Invest, Curiosity Venture Capital, and Crowberry Capital participating — about $67M total since 2018. ARR more than doubled year over year on flat headcount of roughly 50 people. Named customers include Finastra, Clio, Cognism, Oyster, Turing, Sendoso, Gorgias, ElevenLabs, and BlackLine.

Why it shows up in RevOps stacks

• The attribution model is inspectable, and that is the whole pitch. Rule-based models ship out of the box — first-touch, last-touch, linear, U-shaped, W-shaped — alongside a data-driven model that derives credit from your own journeys instead of a fixed rule. W-shaped is the documented default for reporting because it weights the first touch, the lead-creation touch, and the opportunity-creation touch rather than spreading credit evenly. Custom models let you exclude specific session types from receiving credit or set a hierarchy of attribution groups with a linear fallback. When finance asks why a channel got credit, the answer is a rule you wrote, not a black box.
• Funnel definitions are yours. Dreamdata calls a funnel goal a Stage Model, and you define the stages against your own CRM object states. This is what stops the classic attribution failure where the vendor’s idea of “opportunity” and your idea of “opportunity” differ by one stage and every number is quietly wrong.
• Dreamdata AI ships as three separable pieces, launched September 2026. The Analytics Agent answers plain-English questions in-app and shows the filters, model, and date range used to build the report — the audit trail is the feature. The MCP Server puts the same data inside Claude or another LLM client. The Data Warehouse add-on exports the account-based model as a structured schema to BigQuery, Amazon S3, Snowflake, or Azure Storage so your own agents read a modeled schema instead of a raw event lake. Buy one, two, or all three.
• The free tier is real, not a trial. $0 gets B2B web analytics, cookie and cookieless tracking, company identification, engagement scoring, the audience builder, ad spend reporting, and Slack/Teams notifications, capped at 5 seats, 2 months of user history, 3 stage models, 2 notifications, and 1 sync. That cap is tight enough to force an upgrade at production scale and generous enough to prove the identification and journey model on your own traffic before a single procurement conversation.
• Integration coverage is broad on the sources that matter for attribution. Salesforce, HubSpot, Microsoft Dynamics, and Pipedrive on CRM; Marketo, Pardot, Eloqua, and Salesforce Marketing Cloud on automation; LinkedIn, Google, Meta, Microsoft, and Reddit on paid, plus G2, Capterra, and TrustRadius on review traffic; 6sense, Demandbase, RollWorks, ZoomInfo, Clearbit, and Dealfront on intent; Outreach, Salesloft, Apollo, and Chili Piper on sales.

Pricing

Three plans: Free at $0/month, Starter, and Advanced. Only the free tier carries a published price — Starter and Advanced are quote-only on the pricing page as of 4 September 2026, and the $750/month “Activation Starter” figure that circulates in third-party pricing roundups is not on the vendor’s page today. Treat it as stale until a quote confirms it. The Data Warehouse is a paid add-on on top of a plan, not a plan feature.

Vendr’s marketplace data, last updated February 2026, puts the median Dreamdata buyer at $35,440 per year with a range of $27,000 to $96,000. Deployment size moves the number more than seats do: under 5,000 tracked accounts lands around $15,000-$28,000, 5,000-20,000 accounts around $25,000-$50,000, and 20,000+ accounts at $50,000-$75,000 and up. Contracts are annual; monthly billing is not offered to new customers.

Against the alternative: HockeyStack’s reported median annual contract sits near $28,000 with entry pricing around $1,399/month and no free tier at all, so Dreamdata’s landing zone is comparable at the median but the entry point is $0 rather than a quote. Factors.ai starts at $399/month with a free tier and undercuts both, with multi-touch attribution gated behind its Growth tier.

Best for

RevOps and demand-gen leaders at $20M-$200M ARR B2B companies running paid, events, and outbound into multi-stakeholder buying groups, where the binding requirement is an attribution number finance will accept and the team already owns a warehouse it wants the model landed in. It is also the correct first pick for anyone who wants to validate account-level attribution on real traffic before spending budget — no other vendor in this category lets you do that for free.

Pick something else when the shape is wrong. Choose HockeyStack if the reason you are buying is a prebuilt agent layer that acts on the model — forecast challenging, win/loss analysis, closed-lost resurrection — rather than an attribution model you query yourself; it is the other required entry in this category and the closest structural substitute. Choose Adobe Marketo Measure (formerly Bizible) if you are standardized on Adobe and Marketo and attribution can ride an existing enterprise agreement instead of becoming a new vendor. Choose Factors.ai if you are under $10M ARR with a LinkedIn-led ABM motion and a budget that cannot absorb a $27,000 floor. Choose RevSure if the question is forward-looking — will this quarter’s pipeline convert — rather than backward-looking credit assignment.

Watch-outs

• Two of the three plan prices are quote-only, and the add-on is priced separately. Teams model the budget off the free tier and the $750 figure floating around in pricing blogs, then get a quote anchored to account volume. Guard: get Starter, Advanced, and the Data Warehouse add-on quoted as three separate line items in the first pricing conversation, ask where the tracked-account tier boundaries sit in writing, and benchmark against the Vendr $27,000-$96,000 band before you counter.
• The MCP server is text-to-SQL over BigQuery, not a governed semantic layer. It requires the Data Warehouse add-on, authenticates through an OAuth app you create in your own Google Cloud project with MCP Tool User, BigQuery Job User, and BigQuery Data Viewer roles, and the agent infers meaning from the schema. Vendor documentation says so directly: results are non-deterministic and need verification. Guard: keep board-facing and finance-facing numbers on the in-app Analytics Agent, which shows its filters and model, and scope the MCP path to exploration. Build a fixed set of 10-15 reference queries with known answers and re-run them after every schema change.
• Attribution output is capped by CRM hygiene, and no vendor fixes that for you. Inconsistent opportunity stages, missing closed-won dates, and undisciplined UTMs produce a confident model of the wrong funnel. Guard: define your Stage Models against audited CRM stage definitions before the trial starts, and run 30 days in parallel with your existing reporting — if the two disagree, resolve the source data before either number reaches finance.
• Free-tier history is 2 months, which is shorter than most B2B sales cycles. A 6-month enterprise cycle cannot be evaluated on a 2-month window, so the free tier proves tracking and identification, not attribution accuracy. Guard: use free to validate company identification rates and journey stitching against a list of accounts you already closed, and plan on a paid tier plus a backfill conversation before you judge the attribution model itself.

For the deal-inspection and forecasting layer downstream of this, see Gong and Clari; for the intent sources upstream, see 6sense, Demandbase, and Common Room.