heyarnoux.

The model

What the bowtie funnel gets right, and where it breaks in practice

The people who built the bowtie call it two things: a customer journey model and a data model. Almost every team that adopts it takes the first and quietly skips the second, and that is the whole story of why it works on a slide and stalls in a quarterly review. This is not another explanation of the diagram. It is what the model gets right, the three places it breaks once you try to populate it, and the build that decides which of those happens to you.

There are already good explanations of the shape, and most of the results around this one are explanations of the shape. If you have seen the diagram and want to know whether it will survive contact with your own data, that is the question this page takes.

300+companies in Winning by Design's own
Bowtie benchmarking dataset
2,000+go-to-market executives certified
in its Revenue Architecture programme
Two modelswhat the Bowtie is called by its authors:
a customer journey model and a data model
Winning by Design press release on the updated Bowtie Model and Benchmarks, 11 October 2023, read 24 September 2026. The benchmark dataset and the certification count are Winning by Design's own, describing its own tool and its own programme

What the model actually claims

The bowtie takes the familiar narrowing funnel, puts a knot where the deal closes, and mirrors a second shape out to the right for what happens afterwards. The name is a description of the picture. The claim underneath it is stronger than the picture suggests, and it is worth reading in the original:

"A Customer Journey Model and Data Model that extends the traditional marketing and sales funnel to better represent the recurring revenue and recurring impact journey of SaaS customers."Winning by Design, on the updated Bowtie Model, 11 October 2023

Read that as two separate promises, because it is two separate promises. A customer journey model is a shared way of describing what happens to a customer. A data model is a specification for what you store, keyed how, so the description can be checked against reality. The first is a conversation. The second is a schema. Winning by Design's own framing of the update says as much:

"The Bowtie helps GTM teams better describe the customer journey, especially after the initial sale, and provides a common language and data model to optimize this journey for impact and costs."Dominique Levin, CEO of Winning by Design, 11 October 2023

What it gets right, and it is not a small thing

The traditional funnel stops measuring at the signature. In a business where revenue recurs, that is the precise moment before everything that determines whether the deal was worth doing. Levin's argument for the model is economic and it holds:

"Retention and expansion of existing customers takes center stage, as this revenue is less 'expensive' than revenue secured by landing new customers."Dominique Levin, CEO of Winning by Design, 11 October 2023

Acquisition cost in a recurring business is recovered across the customer lifetime, not at the point of sale. A model that ends at closed-won therefore measures the expensive half in detail and the half that pays for it not at all. Giving onboarding, adoption and expansion the same status as awareness and selection is a genuine correction, and the reason the model spread as fast as it did.

The second thing it gets right is quieter. It gives the two halves different units. The left half is about speed, getting to a decision without wasting the pipeline. The right half is about duration, how long the revenue survives and how much it grows. Those are not the same quantity and most funnel thinking treats them as though they were.

Where it breaks, and it is the same place every time

Three failures, in rough order of how often they show up.

1. The right half needs a join that nobody in go-to-market owns

The left half of the bowtie can be populated out of the CRM, because everything it measures was already being recorded there by people whose job is to record it. The right half cannot. Adoption lives in the product database. Money lives in billing. The account lives in the CRM. The right half of the bowtie is a question about all three systems at once, and those systems were built at different times by different teams with different notions of what the unit of a customer is.

So the honest first task is not drawing the model. It is deciding what a customer is, in a way three databases can agree on, and that is an identity problem before it is a reporting problem.

2. The diagram gets adopted and the data model does not

This is the common one. A team runs the workshop, redraws its funnel as a bowtie, agrees the right half matters, and then reports on it with whatever the CRM already had. The stages on the right get populated with proxies: a renewal date, a support ticket count, a health score somebody assigns by feel. The picture updates and nothing underneath it changed.

You cannot draw your way to a data model. The bowtie is a schema wearing a diagram, and most teams hang the diagram.

The failure is hard to see from inside, because the artefact looks more rigorous than what it replaced. A funnel with six stages and no instrumentation at least looks unfinished. A bowtie with proxies on the right looks complete and reports numbers in meetings.

3. It is scoped to recurring revenue, by its own definition

Winning by Design is explicit that the model describes "the recurring revenue and recurring impact journey of SaaS customers". That is a boundary, stated by the authors, and it gets dropped in the retelling. A business selling one-off implementations, hardware or projects can borrow the shape, but the economic argument that justifies the right half does not transfer, because there is no lifetime over which the acquisition cost amortises. Borrowing the diagram without the economics gives you a longer funnel, not a better model.

Worth saying plainly about the benchmarks as well: the 300-company dataset is Winning by Design's own, gathered through its own tool, and the 2,000 certified executives are certified by Winning by Design. That does not make the figures wrong. It makes them a vendor's dataset rather than an independent one, and they are frequently quoted as though they were the latter.

What a GTM engineer actually builds for it

If you want the right half to produce numbers somebody will defend, the work is plumbing rather than strategy. Four layers, and the failure mode in each.

LayerWhat it holdsWhere it breaks
CRMThe commercial record, keyed on an accountOne account, several product workspaces, after any self-serve signup or acquisition
ProductAdoption events, keyed on a user or a workspaceEvents frequently carry no account identifier at all, so usage cannot be rolled up
BillingSubscriptions and the money, keyed on a subscriptionSubscription boundaries rarely match CRM account boundaries
Identity resolutionThe spine that joins the other threeOwned by nobody, budgeted by nobody, and the reason the model stalls

The sequencing that works is narrower than the diagram invites. Pick one metric on the right half, instrument it end to end, and let it be correct before adding a second. One defensible number about adoption is worth more than six stages of proxies, because the first can be acted on and the second can only be presented.

This is ordinary data engineering applied to revenue, which is why the seat that does it ended up with the name it has. The GTM engineer explainer covers where that role came from, and the day-to-day version of the work covers what the same discipline looks like inside a tool.

When the bowtie is the wrong model

  • When the revenue does not recur. The authors scope it to recurring revenue. Without a lifetime, the right half has no economic argument behind it.
  • When the constraint is genuinely on the left. If too few qualified opportunities arrive, redrawing the funnel does not produce any, and the right half is a distraction from a problem that is already visible.
  • When nobody will own the join. Adopting the model without the instrumentation produces a diagram everyone agrees with and numbers nobody can check. That is worse than the funnel it replaced, because it carries more authority.
  • When it is being used to relabel rather than to measure. Renaming customer success as the right half of the bowtie changes the org chart's vocabulary and nothing else.

The sequencing questionDeciding which business unit gets rebuilt first, and in what order, is a separate problem from choosing a revenue model. The playbook for that is written up on its own page.

GTM transformation, one BU at a time

Questions people ask

What is the bowtie funnel?

The bowtie funnel is a revenue model that extends the traditional marketing and sales funnel past the closed deal, so the work after the sale gets the same treatment as the work before it. It is named for its shape: the familiar narrowing funnel on the left, a knot at the point the deal closes, and a second shape widening out to the right covering onboarding, adoption and expansion. Winning by Design, which created it, describes it as a Customer Journey Model and Data Model that extends the traditional marketing and sales funnel to better represent the recurring revenue and recurring impact journey of SaaS customers. Two things in that description are usually missed. It is a data model as well as a diagram, and it is scoped to recurring revenue by its own wording.

Who created the bowtie funnel?

Winning by Design, the B2B revenue consultancy founded by Jacco van der Kooij. The company released an updated version of the model along with a benchmarking tool on 11 October 2023, positioning it as a common language and data model for recurring revenue businesses. At that release it said the Bowtie Benchmarking Tool dataset had surpassed 300 companies and that over 2,000 go-to-market executives had been certified in its Revenue Architecture programme. Both figures come from Winning by Design's own announcement and describe its own dataset and its own certification, which is worth knowing when the benchmarks are quoted as though they were independent industry numbers.

How is the bowtie funnel different from a traditional sales funnel?

A traditional funnel ends at the closed deal, which means every stage it measures is a stage that happens before any money recurs. The bowtie treats the closed deal as the midpoint rather than the finish, and gives onboarding, adoption and expansion the same status as awareness and selection. The argument behind that is economic rather than aesthetic. Winning by Design's CEO Dominique Levin put it as retention and expansion of existing customers taking centre stage, as this revenue is less expensive than revenue secured by landing new customers. In a recurring-revenue business the acquisition cost is only recovered across the customer lifetime, so a model that stops measuring at the signature stops measuring immediately before the part that decides whether the acquisition was worth it.

What data do you actually need to use the bowtie model?

Three systems that were never designed to agree with each other. The commercial record lives in the CRM, keyed on an account. Adoption lives in the product database, keyed on a user or a workspace and usually carrying no account identifier at all. Money lives in the billing system, keyed on a subscription. The right half of the bowtie is a question about all three at once, so using the model properly means resolving those identities into one spine before any stage conversion can be calculated. This is the work that gets skipped, because it is unglamorous, it belongs to nobody by default, and the diagram looks finished without it. A bowtie drawn on a slide costs an afternoon. A bowtie you can query is an integration project.

When is the bowtie funnel the wrong model to use?

When the revenue does not recur, when the constraint is on the left half anyway, or when nobody will own the instrumentation. Winning by Design scopes the model to the recurring revenue and recurring impact journey of SaaS customers in its own definition, so a business selling one-off projects or hardware is borrowing a shape rather than applying the model. If the honest problem is that too few qualified opportunities arrive, the right half is not where the answer is, and redrawing the funnel will not produce one. And if no team will own the join between product, billing and CRM data, adopting the model produces a diagram everybody agrees with and no numbers anybody can check, which is worse than the funnel it replaced because it looks more rigorous.

Sources

  • The Bowtie described as a Customer Journey Model and Data Model extending the traditional funnel to represent the recurring revenue and recurring impact journey of SaaS customers; the Dominique Levin quotations on retention, expansion and a common language and data model; the Jacco van der Kooij quotation; the Bowtie Benchmarking Tool dataset surpassing 300 companies; and over 2,000 go-to-market executives certified in Revenue Architecture · Winning by Design, Updated Bowtie Model and Benchmarks, 11 October 2023, read 24 September 2026
  • Note on what is not sourced here: Winning by Design's 11 October 2023 announcement does not enumerate named stages of the customer journey, and no stage labels or lettered conversion markers are asserted on this page as a result. The businesswire copy of the same release returns HTTP 403 to automated requests, so it was read on winningbydesign.com

The half nobody instruments

Want the right half to produce numbers you can defend?

Most teams asking about the bowtie already have the diagram. What is missing is the join between product, billing and CRM that makes the right half queryable. Tell me which system matters most and you get example profiles and a call with your pick before anything starts.

Placement runs through AYTalent. Engineers are employed and trained on the bench before they start with you.