Funnel
A governed marketing data layer that collects, models, and exports ad data anywhere
Funnel is a marketing data platform that pulls spend, impression, and conversion data out of hundreds of advertising, analytics, ecommerce, and CRM sources, harmonizes it into one consistent schema with its own currency conversion and business rules, and then exports it to dashboards, spreadsheets, BI tools, or a cloud data warehouse. It is sold as a data foundation rather than a reporting tool: the visualization layer is optional, and most buyers use Funnel to feed Looker Studio, Power BI, Tableau, BigQuery, or Snowflake. Pricing starts at roughly $300 per month billed annually, with a usage meter called Flexpoints on top of the plan fee.
Overview
Most reporting tools in the agency category start from the deliverable: a client dashboard, a branded PDF, a scheduled email. Funnel starts a layer below that, from the observation that the hard part of marketing reporting is not drawing the chart but getting sixteen advertising APIs to agree on what a day is, what a currency is, what a campaign is called, and whether a conversion counted. Funnel's product is that reconciliation work, packaged so a marketing team does not have to hire a data engineer to do it.
The platform is organized around Data Hub, the collection and modeling layer. Connectors authenticate against the ad platforms and pull historical and ongoing data, including automatic backfilling when a platform restates figures after the fact. Funnel then applies a shared schema: a base layer of standardized fields plus a semantic layer where you define what your organization means by paid social, by revenue, by market. Dimensions can be parsed out of campaign naming conventions with rules, so a campaign called uk_brand_search_q3 becomes a country, a channel, and a quarter without anyone editing a spreadsheet. Costs get converted to a reporting currency at the rate you choose. The result is one clean table that every downstream tool reads.
Above Data Hub sit two other pieces. Funnel Dashboards and the reporting exports handle presentation, either natively or by feeding Looker Studio, Google Sheets, Excel, Power BI, and Tableau; Funnel MCP exposes the same governed data to Claude, ChatGPT, and other MCP clients so an analyst can ask questions in natural language against real numbers. Funnel Measure, built on the 2024 acquisition of the German attribution firm Adtriba, adds marketing mix modeling, multi-touch attribution, and incrementality testing. Measure is an add-on and requires a Business or Enterprise plan.
The honest positioning caveat: Funnel is at the expensive end of this category and has moved further that way. The free plan was removed in early 2026 and the entry tier now lists at roughly $300 per month billed annually, with a 14-day trial as the only way to try it without committing. Enterprise pricing is quote-only. A three-person agency reporting on five clients will find cheaper tools that do the visible half of the job; Funnel earns its price when the number of sources, currencies, markets, and downstream consumers has grown past what a connector-plus-spreadsheet setup can hold together.
Best for
Larger agencies, in-house performance teams, and multi-market advertisers with many data sources and more than one downstream consumer of the numbers: the buyer who needs one governed definition of spend and revenue feeding a warehouse, a BI tool, and a client dashboard at the same time.
Not the right fit for
- Small agencies with a handful of clients on two or three ad platforms; the plan fee alone exceeds what a purpose-built client reporting tool costs, and most of what you pay for is modeling you do not yet need.
- Teams that want finished, white-labeled client reports out of the box; Funnel's dashboards exist but presentation is not where the product is strongest, and many customers pair it with a separate reporting front end.
- Buyers who need a predictable flat bill; the Flexpoints usage meter sits on top of the plan and is the most frequently cited source of billing surprise.
- Anyone who wants to evaluate for free indefinitely; the free plan was discontinued in early 2026 and only a 14-day trial remains.
- Organizations that only need a warehouse pipeline and have data engineering in house; a general-purpose ELT tool moves the same rows for less, without the marketing semantic layer.
How it works
- 1
You connect data sources by authenticating each account: Google Ads, Meta, TikTok, LinkedIn, Amazon Ads, Google Analytics 4, Shopify, HubSpot, and several hundred others, plus file uploads and a generic API and webhook path for anything without a native connector. Funnel pulls full history on connection and keeps re-fetching recent windows so late-arriving conversions and platform restatements are corrected rather than frozen.
- 2
Incoming fields are mapped into a shared data model. Funnel's base layer standardizes the common metrics across platforms so cost, impressions, clicks, and conversions mean the same thing regardless of source. On top of that you build custom dimensions and metrics with rules: parse campaign names into structured attributes, group sources into channels, define blended ROAS once, and apply currency conversion at a fixed or daily rate.
- 3
Data is then shared out. Funnel writes to Google Sheets, Excel, Looker Studio, Power BI, and Tableau for reporting, and to BigQuery, Snowflake, Amazon S3, Azure, and other cloud destinations for teams that want the warehouse to be the source of truth. Exports run on a schedule, and because the model is defined in Funnel rather than in each destination, changing a metric definition updates every downstream report at once.
- 4
Two layers sit on top of the modeled data. Funnel Dashboards provides native charting and client-facing views inside the product, and Funnel MCP lets an AI assistant query the same governed dataset conversationally. Funnel Measure, sold separately, runs marketing mix models, data-driven multi-touch attribution, and incrementality tests against the collected data and returns budget scenarios rather than raw tables.
Feature breakdown
27 features in 5 modulesData collection
The connector layer, plus the maintenance work that makes marketing APIs usable.- 600 plus connectors
- Advertising, analytics, ecommerce, CRM, email, affiliate, and offline sources, with connector availability tiered by plan (roughly 117 on Starter, 579 on Business, 600 plus on Enterprise).
- Automatic historical backfill
- Connecting an account pulls its full available history rather than starting from today, which matters when you need year-over-year comparisons on day one.
- Restatement handling
- Funnel re-fetches recent date windows so attributed conversions that a platform revises days later are corrected in your data instead of leaving a stale number in the report.
- File and API imports
- CSV and spreadsheet uploads, scheduled file fetches, and a generic API and webhook connector cover offline sales, budgets, and any platform without a native integration.
- Connector monitoring and alerts
- Broken authentications and failed pulls are surfaced with alerting, which is the difference between a client noticing a gap and you noticing it first.
Modeling and governance
Where Funnel differs from a connector tool: one definition of every metric, applied everywhere.- Standardized base layer
- Common fields across platforms are normalized on ingest so cost, clicks, impressions, and conversions align without per-source mapping work.
- Semantic layer
- Business definitions live in Funnel: what counts as paid social, what revenue means, which markets roll up to which region, defined once and inherited by every export.
- Rule-based dimension parsing
- Naming conventions in campaign, ad set, and ad names are parsed into structured dimensions with pattern rules, turning a naming standard into queryable fields.
- Custom metrics and calculations
- Blended ROAS, cost per qualified lead, contribution margin, and similar derived metrics are built as formulas over the shared model rather than in each dashboard.
- Multi-currency conversion
- Spend in local currencies converts to a chosen reporting currency using fixed or dated exchange rates, an unglamorous feature that is decisive for multi-market advertisers.
- Data quality guarantee
- Business and Enterprise plans carry a contractual data guarantee covering connector accuracy and availability, which higher-stakes buyers treat as the reason to move up a tier.
Exports and destinations
Getting the modeled data to wherever the reporting actually happens.- Spreadsheet exports
- Scheduled writes to Google Sheets and Excel, still the most common last mile for marketing reporting no matter what the vendor slide says.
- BI tool connectors
- Looker Studio on every plan, with Power BI and Tableau connectors unlocked on Business and above.
- Cloud warehouse destinations
- BigQuery and Amazon S3 from the Business tier, Snowflake on Enterprise, plus Azure and other cloud targets, with scheduled incremental writes.
- Destination count by tier
- Roughly 13 export destinations on Starter, 46 on Business, and 47 on Enterprise, so which downstream tools you use effectively selects your plan.
- Shared definitions downstream
- Because metrics are defined in Funnel rather than in each destination, a change to a definition propagates to every report instead of requiring edits in five places.
Reporting, AI, and measurement
The layers above the data foundation, including the parts sold separately.- Funnel Dashboards
- Native dashboards and charts built on the modeled data, available on all plans, usable for internal review or client-facing views.
- Funnel MCP
- A Model Context Protocol server that lets Claude, ChatGPT, and other MCP clients query your governed marketing data conversationally, included from the Starter plan.
- Funnel AI
- Assistive features for building connectors, writing transformation logic, and surfacing anomalies in the collected data.
- Marketing mix modeling
- Part of the Funnel Measure add-on: statistical models that estimate channel contribution from aggregate spend and outcome data, without relying on user-level tracking.
- Multi-touch attribution
- Data-driven MTA inherited from the Adtriba acquisition, built on neural network models over path data rather than fixed first or last touch rules.
- Incrementality testing
- Geo and holdout style experiments run against the modeled data to check whether the model's channel estimates survive contact with a controlled test.
- Budget scenario planning
- Measure returns allocation scenarios rather than only historical attribution, which is the output a CMO can actually act on.
Administration
Workspaces, access, and the controls that matter once more than one team is inside.- Unlimited users and workspaces
- From the Business tier, seats and workspaces are uncapped, so cost scales with data usage rather than headcount, which suits agencies with many staff on few sources.
- Roles and permissions
- Advanced user management with role assignment and audit logs on Enterprise, scoping which teams or clients each person can see.
- Enterprise SSO and provisioning
- SAML, OIDC, and SCIM provisioning are Enterprise-tier features, so smaller plans manage accounts manually.
- Flexpoints usage monitoring
- A usage bar in the product shows current Flexpoint consumption, which is a feature you will use often given that consumption drives the bill.
Use cases
4 documentedPerformance agency reporting across 40 client accounts
Every client has a different mix of Google, Meta, TikTok, and Amazon accounts, campaign naming is inconsistent between account managers, and the monthly reporting cycle consumes several days of manual spreadsheet work.
Connectors and parsing rules turn naming conventions into structured dimensions once; a single modeled dataset feeds a Looker Studio template per client, and month-end reporting becomes a review pass instead of a rebuild.
In-house team at a multi-market retailer
Spend sits in eleven currencies across six markets, finance and marketing quote different numbers for the same quarter, and the argument recurs at every review.
Currency conversion and a shared semantic layer produce one reconciled figure that both the BI dashboard and the finance export read from, ending the reconciliation meeting rather than automating it.
Data team building a marketing warehouse
Engineering has a warehouse but no appetite for maintaining a dozen brittle marketing API integrations that break every time an ad platform revises its schema.
Funnel handles collection and normalization and writes clean incremental tables into BigQuery or Snowflake; the data team owns the modeling downstream and stops fielding connector breakages.
CMO deciding next year's channel budget
Platform-reported conversions add up to more revenue than the company actually booked, and nobody can say what an incremental dollar in a given channel is worth.
Funnel Measure runs a mix model on the collected data, calibrated against a geo holdout test, and returns allocation scenarios with confidence ranges instead of a stack of self-reported platform ROAS.
Pricing
from About $300 per month on Starter, billed annuallyTwo components: a plan fee that determines which connectors, destinations, and features you can use, plus Flexpoints, a usage meter consumed by the data sources you connect, the volume you pull, and the transformations you apply. Plans are quoted billed annually. Funnel Measure is a separate add-on requiring Business or Enterprise.
| Plan | Price | Includes |
|---|---|---|
| Starter | About $300 per month, billed annually |
No cloud warehouse destinations and no Funnel Measure. This is the tier for a single team feeding spreadsheet and Looker Studio reports. |
| Business | About $600 per month, billed annually |
The practical starting point for agencies and anyone writing to a warehouse. Funnel Measure can be added here. |
| Enterprise | Custom quote contact sales |
|
Add-ons
- Funnel Measure (Quoted separately): Marketing mix modeling, data-driven multi-touch attribution, and incrementality testing. Requires a Business or Enterprise plan and is not sold standalone.
- Additional Flexpoints (Quoted per allocation): Every plan includes a base allocation; connecting more sources, pulling more data, or adding destinations consumes more, and running out means buying more or disconnecting something.
Billing notes
- Funnel removed its free plan in early 2026 and revised list pricing more than once during that year, so any figure older than a few months should be re-checked against the pricing page before it goes in a budget.
- The Flexpoint meter, not the plan fee, is what makes bills unpredictable; consumption rises as you add accounts, sources, and export destinations, and customers regularly report the effective cost drifting well above the headline plan price.
- Flexpoints are released when you disconnect a source, so an unused connector left authenticated is a live cost. Auditing connections periodically is part of owning the tool.
- Which downstream tools you use effectively picks your tier: Power BI, Tableau, BigQuery, and S3 are Business features, and Snowflake is Enterprise only.
- Contracts are annual and several customers describe notice periods and auto-renewal terms as inflexible, so read the renewal clause before signing.
- The 14-day trial runs with unlimited Flexpoints, which means it is useful for testing capability but does not reveal what your steady-state usage will actually cost.
Value assessment: Funnel is priced as infrastructure, and it is worth infrastructure money only when the data problem is real: many sources, several currencies or markets, and more than one downstream consumer that must agree on the numbers. In that situation the semantic layer is genuinely difficult to replicate and cheaper tools do not attempt it. Below that threshold the value is poor. An agency with five clients on three platforms is paying enterprise plumbing rates for a job a connector tool plus a reporting front end does for a fraction of the price, and the Flexpoint meter adds cost variance that small firms tend to price badly. The honest test is whether anyone in your organization currently spends days reconciling marketing numbers by hand; if not, Funnel is over-specified for you.
Strengths & limitations
Strengths
- The strongest data modeling layer in marketing reporting: business definitions, derived metrics, and dimension parsing live in one place and propagate to every destination.
- Connector breadth and maintenance quality, including historical backfill and re-fetching of restated platform figures rather than freezing whatever was true on pull day.
- Multi-currency handling that is a first-class feature rather than a spreadsheet workaround, which is decisive for multi-market advertisers.
- Vendor-neutral about presentation: the same modeled data feeds Sheets, Looker Studio, Power BI, Tableau, and a warehouse, so you are not locked into the vendor's charting.
- Funnel MCP exposes governed data to AI assistants, which is a materially better answer than letting a chatbot query raw ad platform APIs.
- Funnel Measure brings mix modeling and incrementality testing into the same platform as the raw data, avoiding a separate data pipeline for the measurement vendor.
- Unlimited users and workspaces from the Business tier, so cost scales with data rather than headcount.
Limitations
- The Flexpoints meter makes bills genuinely hard to forecast; users consistently report costs climbing well beyond the plan fee as sources and destinations accumulate.
- No free plan since early 2026, and only a 14-day trial with unlimited Flexpoints, which tests features but not what your real usage will cost.
- Entry pricing of roughly $300 per month billed annually puts it out of reach for small agencies that a cheaper client-reporting tool serves adequately.
- Enterprise pricing is quote-only, and Snowflake, SSO, SCIM, and audit logs all sit behind that unpublished number.
- Connector counting can be unintuitive: separate tabs of the same Google Sheet, or separate ad accounts under one login, can each consume allocation.
- Native dashboards are functional rather than a strength; teams that want polished white-labeled client reports usually add a second tool for presentation.
- The learning curve is real for non-analysts. The semantic layer is the product's advantage and also the part that requires someone who thinks in data models, with reviewers noting limited in-product guidance.
- Funnel Measure is an add-on gated behind Business or Enterprise, so mix modeling is not available to entry-tier buyers at any price.
Head-to-head comparisons
5 alternativesFunnel vs Improvado
from Free Limited at $0 per month; MCP Only at $100 per month; Advanced and Enterprise by quoteThe closest direct competitor: both sell a governed marketing data layer with heavy transformation rather than a reporting front end, and both land in the same enterprise price band. Funnel is more self-serve and more transparent about list pricing, with a published Starter tier and a 14-day trial. Improvado leans harder on managed onboarding and custom pipeline work, which suits buyers who want the vendor to build the model for them. If you have an analyst who will own the semantic layer, Funnel is usually the faster path; if you want the work done for you, Improvado's services model is the argument.
Full Funnel vs Improvado comparisonFunnel vs Cyfe
from $29 per month (Starter: 2 dashboards, 1 user)Barely the same category. Cyfe is an inexpensive dashboard product aimed at small businesses that want widgets on a screen, with essentially no modeling layer and no warehouse export. Funnel is a data platform that costs an order of magnitude more and does a fundamentally different job. Choose Cyfe if the requirement is a live dashboard for a small stack; choose Funnel if the requirement is that everyone agrees what the numbers mean before they are charted.
Full Funnel vs Cyfe comparisonFunnel vs Supermetrics
from €49 per month (Starter), or €39 per month billed yearlyThe most common head-to-head in this category. Supermetrics is the pragmatic choice for pulling ad data into Sheets, Looker Studio, or a warehouse, with per-connector pricing that is easier to predict and a lower entry point. Funnel's advantage is everything that happens between extraction and reporting: the shared schema, custom dimensions from naming rules, currency conversion, and definitions that propagate downstream. Teams that only need the rows moved should buy Supermetrics; teams that need the rows reconciled should look at Funnel.
Full Funnel vs Supermetrics comparisonFunnel vs Databox
from $0 (Free), then $64 per month billed annually (Analyst)Different halves of the same workflow. Databox is a dashboard and goal-tracking product with a genuinely usable free tier and a strong mobile and alerting story; its modeling is shallow by design. Funnel does the modeling and expects you to bring your own presentation layer. A few teams run both, with Funnel feeding a warehouse and Databox reading a curated subset, but for a small business choosing one, Databox costs a fraction and answers the visible question.
Full Funnel vs Databox comparisonFunnel vs Fivetran
from Free for up to around 500,000 monthly active rows; paid usage from roughly $500 per month at modest volumesFivetran is general-purpose ELT: it moves data from hundreds of source systems into a warehouse reliably and leaves all modeling to dbt and your data team. Funnel is marketing-specific and ships the modeling with it, including currency conversion and campaign-name parsing that a general ELT tool has no opinion about. Organizations with a data team and an existing warehouse stack often find Fivetran plus dbt cheaper and more flexible; marketing teams without that engineering support get far more out of the box from Funnel.
Full Funnel vs Fivetran comparisonImplementation & onboarding
- Setup time
- Connecting the first sources takes an afternoon, and historical data lands automatically. Building the semantic layer, the part that produces the actual value, typically takes two to six weeks depending on how many markets, currencies, and inconsistent naming conventions you are reconciling.
- Learning curve
- Steep for anyone who has not worked with a data model before. The connector setup is easy; the rules engine, custom dimensions, and metric definitions reward someone who thinks in schemas, and reviewers frequently note that in-product guidance is thinner than the complexity warrants.
- Onboarding
- Self-serve on Starter with documentation and a knowledge base; Business and Enterprise buyers typically get assisted onboarding and a customer success contact. Enterprise adds an SLA and priority support.
- Migration notes
- Historical data is refetched from the source platforms on connection, so you are not migrating rows so much as rebuilding definitions. Budget the time to translate existing spreadsheet logic into Funnel rules; that translation, not the data movement, is the migration. Before cancelling, export the modeled data you need, since the platform is the layer that reproduces it.
Platform, API & security
- Platforms
- Web applicationCloud data warehouse destinationsBI tool connectorsModel Context Protocol server
- API
- REST API for programmatic access to workspaces, connections, and exports, plus scheduled export jobs to spreadsheets, BI tools, and cloud storage. Funnel MCP exposes the modeled dataset to Claude, ChatGPT, and other MCP-compatible clients.
- Compliance
- GDPRSOC 2 Type IIISO 27001
- Data residency
- Swedish company with EU data processing; specific residency arrangements are handled at the Enterprise tier.
- SSO
- SAML and OIDC single sign-on with SCIM provisioning, available on Enterprise plans only.
- Security notes
- Role-based access control and audit logs are Enterprise features. Because Funnel stores collected marketing data rather than acting as a pass-through, the retention and access terms in the contract matter more than they would for a simple connector tool.
Support & resources
- Channels
- Email supportIn-app chatCustomer success contact on higher tiersEnterprise SLA and priority support
- Documentation
- A detailed knowledge base covering connectors, the rules engine, and export configuration, though the depth of the modeling layer means documentation is often consulted rather than skimmed.
- Community
- An active customer base among European agencies and in-house performance teams, plus a substantial published blog on marketing measurement that functions as category education.
Company
- Founded
- 2014
- Headquarters
- Stockholm, Sweden
- Ownership
- Independent, venture-backed
- Founders
- Fredrik Skantze, Per Made
- Employees
- ~350 (est. 2026)
- Funding
- Over $130 million raised across a 2017 Series A led by Balderton Capital, a 2020 Series B, and a 2021 pre-IPO round.
Funding history
| Round | Amount | Year | Notes |
|---|---|---|---|
| Series A | About $10 million (EUR 8.3m) | 2017 | Led by Balderton Capital. |
| Series B | About $47 million (EUR 42.6m) | 2020 | |
| Pre-IPO | About $66 million (EUR 57.1m) | 2021 |
Timeline
- 2014Founded in Stockholm by Fredrik Skantze and Per Made, previously the creators of the Facebook ads tool Qwaya.
- 2017Raises a Series A led by Balderton Capital and expands beyond the Nordics.
- 2020Series B funding; named Best Marketing Tech Startup at the TechCrunch Europas Awards.
- 2021Pre-IPO round of roughly EUR 57 million; appears on the Financial Times FT 1000 list for the first of four consecutive years.
- 2024Acquires the German attribution company Adtriba and launches Funnel Measure, adding mix modeling and multi-touch attribution to the platform.
- 2025Ships Funnel MCP, exposing governed marketing data to Claude, ChatGPT, and other MCP clients, and repositions around an AI-native data foundation.
- 2026Discontinues the free plan and revises list pricing more than once, settling on a published Starter tier around $300 per month billed annually.
Integrations
- Google Ads
- Meta Ads
- TikTok Ads
- LinkedIn Ads
- Amazon Ads
- Microsoft Advertising
- Google Analytics 4
- Shopify
- HubSpot
- Salesforce
- Klaviyo
- Looker Studio
- Power BI
- Tableau
- Google Sheets
- Google BigQuery
- Snowflake
- Amazon S3
Frequently asked questions
12 questionsWhat is Funnel.io used for?
Collecting marketing data from advertising, analytics, ecommerce, and CRM platforms, harmonizing it into one consistent model with shared metric definitions and currency conversion, and exporting it to spreadsheets, BI tools, or a cloud data warehouse. It is a data foundation rather than a dashboard product, though it includes dashboards.
How much does Funnel cost?
Starter lists at roughly $300 per month billed annually and Business at roughly $600, with Enterprise on a custom quote. On top of the plan fee you consume Flexpoints, a usage meter driven by how many sources you connect, how much data you pull, and how many transformations and destinations you use, which means the effective bill is usually higher than the plan price.
Does Funnel have a free plan?
Not any more. Funnel discontinued its free plan in early 2026. What remains is a 14-day trial of the Starter plan with unlimited Flexpoints during the trial period, which is enough to test the connectors and modeling but will not show you what your steady-state usage costs.
What are Flexpoints?
Funnel's proprietary usage unit. Each plan includes a base allocation, and connected data sources, data volume, transformations, and export destinations consume from it. Disconnecting something releases its Flexpoints immediately. The design lets you mix and match features, but it also makes the monthly cost a moving number rather than a fixed one.
Funnel vs Supermetrics: which should an agency choose?
Supermetrics if the job is moving ad data into Sheets, Looker Studio, or a warehouse: cheaper entry point, per-connector pricing that is easier to forecast, and less to learn. Funnel if the job includes reconciling that data first, meaning shared metric definitions, dimensions parsed from campaign naming conventions, and multi-currency reporting. Most small agencies are better served by Supermetrics; large multi-market ones outgrow it.
Can Funnel export to BigQuery or Snowflake?
BigQuery and Amazon S3 are available from the Business tier. Snowflake is an Enterprise-tier destination. The Starter plan has no cloud warehouse destinations at all, so a warehouse requirement effectively selects your plan before any other consideration does.
Is Funnel a reporting tool or a data pipeline?
Both, but it is bought as a pipeline. The distinguishing work is the modeling layer between extraction and reporting. Funnel Dashboards exist and are usable, though teams that need polished white-labeled client deliverables typically pair Funnel with a dedicated reporting tool rather than relying on the native dashboards.
What is Funnel Measure?
A separately priced add-on covering marketing mix modeling, data-driven multi-touch attribution, and incrementality testing, built on technology from the Adtriba acquisition in 2024. It runs on the data already collected in the platform and requires a Business or Enterprise plan; it cannot be bought on its own or added to Starter.
Is Funnel too expensive for a small agency?
For most, yes. At roughly $300 per month billed annually before Flexpoint consumption, the plan fee alone exceeds what several purpose-built client reporting tools charge for the entire job. Funnel becomes economical when the number of sources, markets, and downstream consumers makes manual reconciliation a recurring multi-day task.
Who owns Funnel?
It remains an independent Swedish company headquartered in Stockholm, founded in 2014 by Fredrik Skantze and Per Made, and backed by investors including Balderton Capital and Eight Roads Ventures. It has not been acquired; it was itself the acquirer of Adtriba in 2024.
How long does Funnel take to implement?
Connecting sources and pulling history takes hours. Building the semantic layer, custom dimensions, and metric definitions that justify the price typically takes two to six weeks, and the work is closer to data modeling than to dashboard configuration. Plan for someone to own it rather than assuming it configures itself.
Does Funnel work with Claude or ChatGPT?
Yes, through Funnel MCP, a Model Context Protocol server included from the Starter plan. It lets an AI assistant query the governed dataset directly, so answers come from your reconciled model rather than from a chatbot guessing at raw ad platform exports. Funnel Measure results are queryable through the same connection.
Editorial verdict
Funnel is the most serious data-modeling product in marketing reporting, and it is priced accordingly. The semantic layer, currency handling, campaign-name parsing, and restatement-aware collection solve problems that cheaper tools in this category do not attempt, and the vendor-neutral export story means you can keep your existing dashboards and warehouse. It is also the least small-business-friendly tool here: the free plan is gone, list pricing starts near $300 per month billed annually, Enterprise is quote-only, and the Flexpoints meter makes forecasting the real bill genuinely difficult, which is the complaint that appears most often in customer reviews. Buy it when marketing numbers are reconciled by hand across many sources and markets and the reconciliation itself has become the bottleneck. If the requirement is client dashboards for a handful of accounts, look further down the price list and revisit Funnel when the data problem is real.
Written by the SaaSTracker editorial team. Awards, when shown, are judged against the published criteria in our methodology.