Skip to main content

Updated Sep 30, 2026

Data models

A data model tells FunnelStory how to turn rows from one of your connected systems into something it understands — an account, a user, a product event, a support ticket, a subscription. Models are the raw-data layer of the Customer Intelligence Graph: predictions, needle movers, the Account Dossier, and every agent reason over what your models bring in.

Connections bring access; models bring meaning​

A connection gives FunnelStory access to a system, such as your warehouse, CRM, or support tool. A model says what to read from it and what each column means: which column is the account ID, which is the renewal date, which is the event name. One connection can feed many models, and some integrations create their models for you — support tickets from Zendesk, for example.

The account model comes first​

Everything in FunnelStory hangs off an account, so the account model is the one model every workspace needs. Other models attach their rows to accounts by account ID, which is why they're configured after it. See Configuring the account model.

Kinds of models​

Models fall into a few families, each answering a different question about an account:

FamilyAnswersFeeds
Who — accounts and usersWhich customers you have, who works there, and who owns them on your sideEverything
What they do — product and non-product activity, account metricsHow the account uses your product, and trends over timePredictions, health scores, Outcome Agents
What they say — conversations, meetings, support tickets, notesWhat customers raise in email, calls, chat, and ticketsNeedle movers, the Account Dossier, relationship maps
What they bought — products, plans, subscriptions, usageContracts, entitlements, renewals, and consumptionRenewal and expansion analysis, Outcome Agents
Your team's work — tasks and strategyFollow-ups and account plansAccount tasks and the Account Strategy tab

The full list of model types, with a reference page for each, is in the Data models overview.

Models stay current on a schedule​

Each model refreshes on an interval — every six hours by default — pulling new and changed rows from its source. Everything computed from the data follows: a new support ticket shows up on the account's timeline and can raise a needle mover, and usage changes flow into health scores and predictions. You can refresh a model by hand when you need data sooner.