How'd you fix Doodle's revenue issues in 2026?

The two options on the table

Every diagnosis of Doodle's revenue issues arrives at the same fork, and most of the wasted effort in a turnaround like this comes from refusing to pick a side of it.

Option A: defend and optimize the consumer base. Doodle's original shape is a free group-poll link you paste into an email thread. Revenue comes from display advertising against enormous, low-intent pageview volume, plus a thin premium tier that removes ads and adds branding. The defense play says: the poll is still the best-known group-scheduling artifact on the open web, the brand has genuine unaided recall in Europe, and the fix is operational — raise ad yield per session, tighten the free-to-paid upgrade prompt, cut infrastructure cost per poll, and improve retention on the premium tier. No new sales org, no new product surface, no new buyer. You are compounding an asset you already own.

The problem is that each input to that model has been moving the wrong way for several years, and none of the movements are Doodle's fault or within Doodle's control. Privacy regulation under GDPR and the ePrivacy regime made consent-gated European ad inventory materially less valuable than the same impressions were a decade ago. Apple's App Tracking Transparency framework, introduced in iOS 14.5, removed the default identifier that made behavioral targeting cheap. Google's own long-running retreat from third-party cookies in Chrome pushed the whole open-web display market toward contextual and first-party signals, where a scheduling utility has little to sell — nobody's purchase intent is legible from "this person is picking a time for a team lunch." Meanwhile the free alternatives multiplied: Microsoft Bookings and FindTime ship inside Microsoft 365, Google's appointment scheduling ships inside Google Workspace, and open-source projects like Rallly and Cal.com give away the poll and the booking link respectively. Defending means grinding out yield improvements on inventory whose clearing price is set by forces above you, against substitutes whose price is zero because they are bundled.

Option B: convert to paid B2B infrastructure. This one says the scheduling *link* is a commodity and the scheduling *decision* is not. The revenue moves from advertising to contracts, and the buyer moves from an individual picking a lunch time to a RevOps or sales-operations leader who owns pipeline throughput and can sign a purchase order. Three engines, sold to three different budget lines:

1. Enterprise meeting-operations contracts, priced in the low tens of thousands to low hundreds of thousands per year, where the deliverable is not seats but measured meeting outcomes — no-show rate, time-to-first-meeting, calendar utilization on customer-facing teams, and cycle-time between stage gates.
2. Vertical packages for small professional-services firms — legal, consulting, healthcare — priced in the hundreds to low thousands per month, where the product ships with the compliance behavior, records retention, and intake logic that the vertical's regulator or malpractice carrier effectively requires.
3. An AI scheduling module sold as an add-on to the first two, which does multi-party consensus using signals a generic assistant cannot see: CRM stage, account tier, rep territory and skill tags, historical no-show behavior per contact, and calendar-fatigue patterns.

The trade-off is unsentimental. Option A preserves the audience and kills the margin. Option B preserves the margin and probably shrinks the audience by an order of magnitude, because the enterprise motion needs product decisions — mandatory accounts, admin consoles, SSO, retention policy, data-processing agreements — that make the frictionless anonymous poll worse. You cannot fully have both. The honest version of the 2026 fix is: run Option B as the growth engine, run Option A in harvest mode with a cost floor, and stop spending product cycles trying to make the free poll monetize like software.

How to decide between them

The decision is not a matter of taste, and it should not be made by whoever argues hardest in the room. It's decidable against four tests, run in order, over about six to eight weeks of instrumented evidence.

Test one — is the ad revenue line decaying structurally or cyclically? Pull thirty-six months of revenue per thousand sessions, segmented by geography and by device, and separate price from volume. If sessions are flat and RPM is falling, that's a price problem set by the ad market and it will not revert. If RPM is flat and sessions are falling, that's a demand problem and it might be fixable with SEO and product work. Structural price decay across both mobile and desktop, in both EU and US inventory, is the signal that Option A is a harvest, not a strategy.

Test two — does any measurable share of the free base look like a business? Instrument the free poll for the shape of a work meeting: participant count above four, weekday business-hours creation, corporate email domains among respondents, and repeat creation by the same organizer within thirty days. If a meaningful slice of active organizers looks like recurring business use, there is a real conversion population to sell into and Option B has warm supply. If the base is overwhelmingly one-off social coordination, Option B has to be sold cold and the go-to-market cost estimate roughly doubles.

Test three — will anyone actually sign? Before any budget reallocation, run ten to fifteen structured discovery calls with RevOps and sales-ops leaders at companies in the 500-to-5,000-employee range, and ask for a paid pilot at real price, not for interest. Two or three signed paid pilots at five figures inside a quarter is a pass. Enthusiastic calls with no purchase order is a fail, and it's the most common way this kind of pivot dies eighteen months later instead of in week six.

Test four — can the balance sheet fund the gap? The B2B motion produces revenue on enterprise timelines. Model the trough honestly: ad revenue declining month over month while new contract revenue ramps from zero, with sales hires carrying cost for two to three quarters before their first close. If the parent company or the cash position cannot absorb four to six quarters of that shape, the sequencing has to change — vertical SMB first, because it closes in weeks rather than quarters, with enterprise following once there are reference logos.

The tests are ordered deliberately. Test one is cheap and can be run from existing data in a week. Test three is the expensive one and should never be run before tests one and two have justified it. A team that starts with test three — hiring a sales leader and hoping — has skipped the part where it learns whether it has anything to sell.

Concrete numbers behind each option

Numbers make the fork legible. These are modeling ranges a practitioner would use to frame the decision, not reported figures for the company.

Option A economics. Ad-supported utility revenue is a function of monthly sessions times ads per session times RPM. Open-web display RPM for non-commercial-intent European inventory sits in the low single digits of dollars per thousand impressions, and a scheduling poll generates few impressions per session because the visit is short and purposeful. Gross margin after ad-tech take rate, consent-management tooling, CDN, and compute lands somewhere in the 30-to-45 percent range. Realistic optimization upside — better ad placement, higher fill, consent-rate improvements, a tighter premium upsell — is perhaps 15 to 25 percent on the revenue line in the first year, and it is a one-time gain against an underlying decline that continues afterward. That's the shape that makes Option A a harvest: real money, no compounding.

Enterprise contract economics. An outcome-tied meeting-operations contract is priced against the labor it protects, not against seat count. A 50-rep customer-facing team losing two hours per rep per week to scheduling friction, at a fully loaded cost of $120 to $175 per hour, is burning roughly $600,000 to $900,000 annually in coordination overhead. A contract in the $40,000 to $150,000 range that credibly removes a quarter to a third of that friction is defensible on arithmetic alone, and it survives procurement review because the buyer can show the model. Expect an average contract value in the mid five figures for a first cohort, gross margin of 70 to 80 percent, and a sales cycle of four to seven months with security review included. Enterprise CAC in this segment realistically lands at $8,000 to $20,000 per closed contract once you load a director, a solutions engineer, and marketing — an order of magnitude above the sub-$50 implicit CAC of an organic freemium signup, which is exactly why it only works at contract prices.

Vertical SMB economics. A three-partner law firm or a twelve-person consultancy will pay $200 to $1,500 per month for scheduling that ships with intake forms, conflict-check prompts, retained records, and calendar behavior matched to their billing model. The cycle is two to six weeks, the buyer is the managing partner or practice administrator, and CAC through content and partner channels lands in the $500 to $2,500 range. At an average of $600 per month, 400 firms is roughly $2.9 million in annual recurring revenue. Gross margin runs 75 to 85 percent. The critical number is logo churn: professional-services SMB software churns 1.5 to 3 percent monthly unless it's embedded in the workflow, so the entire vertical thesis lives or dies on whether the compliance and records features make removal painful.

AI module economics. Sold as an attach to the first two engines, in the $10,000 to $60,000 per year range depending on seat count. Marginal cost is inference plus integration maintenance, so margin lands high — 80 percent and up at scale — but only if inference cost per scheduling decision is metered and capped. The honest risk is that a generic assistant does 70 percent of the job for free next year, so the module has to be priced and positioned around the integrated signals, not the language capability.

The blended picture. Option B at plausible year-two scale — say 30 to 50 enterprise contracts, 300 to 500 vertical firms, and a 30 percent attach rate on the AI module — produces a materially smaller revenue base than a healthy ad business at peak, but at roughly double the gross margin, with revenue that renews rather than needing to be re-earned every session. Net revenue retention is the number that decides whether the pivot was worth it: below 100 percent, you have replaced a declining business with a leaking one. Above 110 percent, the expansion motion inside existing accounts becomes the growth engine and the sales cost stops scaling linearly.

What the comparison sets aside. Neither option makes the competitive picture easy. Calendly is a well-funded independent company that raised at a multi-billion-dollar valuation in 2021 and owns the enterprise scheduling category in North America. Microsoft and Google give away the adjacent functionality inside suites that most target accounts already pay for. That's precisely why Option B is scoped to multi-party consensus, compliance-bound verticals, and European data residency rather than to head-on booking-link competition — those are the places where the bundled free tools are genuinely weaker and where an incumbent's pricing power doesn't reach.

Implementation details and sequencing

The order of operations matters more than any single decision inside it. This is a four-phase sequence spanning roughly twelve to eighteen months, and each phase has an explicit kill gate.

Phase one — evidence and cost floor (weeks 1 to 8). Run the four decision tests. In parallel, put the consumer surface into harvest mode: freeze net-new feature work on the free poll, keep it fast and reliable, reduce infrastructure spend per session, and set a maintenance headcount that does not grow. Do not degrade the free product — it is the top of the funnel and the brand asset. Instrument work-meeting signals so phase two has a target list. The gate: at least one of tests one through three must come back clearly positive, or the pivot is descoped to vertical SMB only.

Phase two — first paid contracts, no infrastructure (weeks 8 to 24). Sell before you build. Take two to four design-partner enterprises through a paid pilot using the existing product plus heavy human service — a solutions person who configures templates, sets up the integration, and reports the outcome metrics manually every two weeks. This is deliberately unscalable, and that's the point: it discovers which outcomes buyers actually care about at a fraction of the cost of guessing in a roadmap. Price the pilot at real money — a discounted-but-genuine five figures — because free pilots teach you nothing about willingness to pay. Simultaneously ship the minimum enterprise-readiness surface, because these deals stall on procurement rather than product: SSO via SAML or OIDC, an admin console with user provisioning, audit logs, a documented data-processing agreement, configurable retention, and a completed security questionnaire package. Budget one to two quarters of engineering for this and treat it as sales infrastructure, not features.

Phase three — verticalize one market, not three (months 6 to 12). Pick legal first. It has the clearest compliance requirements, the most acute pain around client intake and conflict checking, an established willingness to pay for practice tooling, and a well-defined channel through bar associations and practice-management vendors. Ship the template set, the records behavior, and one deep integration with a dominant practice-management system rather than shallow integrations with four. Get to 75 to 150 paying firms and a documented churn number before opening the second vertical. The failure mode here is running legal, consulting, and healthcare concurrently with one product team and shipping three mediocre half-verticals — healthcare in particular carries HIPAA or equivalent obligations that are a program of work, not a template, and it should be last.

Phase four — the AI module and expansion (months 9 to 18). Build the scheduling intelligence layer only after the enterprise accounts exist, because the module's whole defensibility is the account data it reads. Wire it to CRM stage and owner, historical attendance per contact, territory and skill tags, and meeting-density signals. Sell it as an expansion motion into installed accounts, where the CAC is a fraction of new-logo cost and the outcome data from phase two is the proof.

Organizational sequencing. The hires that matter, in order: one enterprise go-to-market lead who has personally closed five-figure-plus SaaS contracts, one solutions engineer who can run pilots and answer security reviews, then a second seller only after the first has closed three deals. Hiring a full sales team before there is a repeatable motion is the most expensive way to learn that the motion isn't repeatable yet. Marketing spend shifts from consumer brand placement toward B2B channels — practitioner communities, vertical trade associations, comparison and review sites, and integration marketplaces — but shift it in tranches tied to pipeline evidence, not all at once.

Governance. Publish a monthly scorecard with five lines: ad revenue, contract ARR, net revenue retention, enterprise pipeline coverage, and free-tier session count. The last one is on the list specifically to catch the failure where enterprise product decisions quietly degrade the free funnel that feeds everything. Any month where free sessions drop more than 10 percent without a known cause, the enterprise roadmap pauses until it's explained. That single guardrail is what keeps the pivot from eating its own top of funnel — the most common way scheduling companies solve their revenue issues on paper and lose distribution in practice.

Related questions

What kills this pivot fastest?

Hiring an enterprise sales team before any paid pilot has converted. The cost carries for two or three quarters against zero contract revenue while ad revenue keeps declining, and the cash trough closes the window before the motion is proven. Sell first, staff second.

Why lead with legal rather than healthcare?

Legal has clear compliance pressure, a reachable buyer in the managing partner, established software budgets, and template-shaped requirements. Healthcare requires HIPAA-grade program work, longer procurement, and integration with clinical systems — real revenue, but a multi-quarter build rather than a beachhead.

Does Swiss data residency actually justify a price premium?

For EU-headquartered enterprises in regulated sectors, yes — it removes a transfer-mechanism review from procurement. But it is a tiebreaker on an otherwise competitive product, not a standalone reason to buy. Price it as a plan tier, not as the entire pitch.

What happens to the free poll?

It stays, fast and unchanged, in harvest mode with a fixed cost ceiling. It remains the brand asset and the discovery surface for work-meeting cohorts. Degrading it to force upgrades destroys the funnel that makes the paid motion cheap.

How do you price an outcome-tied contract without carrying the risk?

Tie a modest slice — 10 to 20 percent of contract value — to a jointly instrumented metric like no-show rate, with a baseline measured during the pilot. Full outcome pricing transfers too much risk onto a vendor with no control over the customer's sales process.

FAQ

Which option should Doodle pick if it can only pick one?

Option B, the conversion to paid B2B infrastructure — but sequenced defensively, with vertical SMB before enterprise if cash is tight. The ad-supported model's price decline is driven by privacy regulation and platform changes that no amount of yield optimization reverses, so optimizing it is a harvest with a fixed ceiling rather than a strategy with compounding upside.

Isn't the free consumer base worth more than a few hundred paying firms?

In revenue at peak, possibly. In enterprise value, no. Contract revenue at 75 to 85 percent gross margin that renews annually is worth a multiple of ad revenue at 30 to 45 percent margin that must be re-earned every session and whose clearing price is set externally. The base still matters — as distribution and brand, not as the monetization engine.

How is the AI scheduling module defensible when general assistants keep improving?

Only through integration, not through language capability. A generic assistant cannot see CRM stage, account tier, territory assignment, per-contact no-show history, or team calendar-load patterns. Build the module on those signals and price it as an add-on to accounts that already supply them. If it's positioned as "smarter time suggestions," it gets commoditized quickly.

What is the single most important number to watch after the pivot starts?

Net revenue retention on contract accounts. Below 100 percent means the new business leaks as fast as it fills and the sales cost scales forever. Above 110 percent means expansion inside existing accounts carries growth, which is what makes an enterprise motion economically superior to the model it replaced.

Can the parent company relationship block this?

It can slow it. A media parent optimized for consumer brand advertising has neither the B2B go-to-market machinery nor the natural instinct to fund a four-to-six-quarter revenue trough. The mitigation is a conservative model presented as a staged bet with explicit kill gates, so each budget tranche is released against evidence rather than as a single large commitment up front.

What does a RevOps buyer actually need to see to sign?

A baseline they trust, a mechanism they understand, and a report they can forward. In practice: current no-show rate and time-to-first-meeting measured from their own calendar and CRM data, a clear explanation of what the product changes about scheduling behavior, and a recurring report tied to those two numbers. Feature lists do not close these deals; instrumented before-and-after does.

Sources

• https://gdpr.eu/what-is-gdpr/
• https://www.edpb.europa.eu/
• https://www.edoeb.admin.ch/edoeb/en/home.html
• https://developer.apple.com/documentation/apptrackingtransparency
• https://privacysandbox.com/
• https://support.microsoft.com/en-us/office/schedule-appointments-with-microsoft-bookings
• https://workspace.google.com/products/calendar/
• https://github.com/lukevella/rallly
• https://github.com/calcom/cal.com
• https://www.iab.com/insights/

Related on PULSE

• [How'd you fix Illinois's NIL & athletic revenue issues in 2026?](/knowledge/q1464)
• [How'd you fix Aston Carter's revenue issues in 2026?](/knowledge/q1480)
• [How'd you fix CyberCoders's revenue issues in 2026?](/knowledge/q1479)
• [How'd you fix Creative Financial Staffing's revenue issues in 2026?](/knowledge/q1478)
• [How'd you fix LanceSoft's revenue issues in 2026?](/knowledge/q1477)
• [How'd you fix Goodwin Recruiting's revenue issues in 2026?](/knowledge/q1476)