On September 13, Fabien Pinckaers posted a 30-page deck to LinkedIn that no one asked him to publish: what happened when Odoo switched off Google Ads entirely.
Software companies rarely, if ever, publish this sort of detail. Churn by acquisition channel is the kind of number that stays in a board deck. Odoo put it on a public feed, where it drew ~1,450 reactions and 90+ comments, and it is the most open look at Odoo's own sales machine I have seen.
My bias, up front: I do not sell Odoo licenses. I depend on Odoo. 19 Prince helps US companies on Odoo drive more leads, opportunities and sales, so the part of this deck I care about is what happens to demand in this market.
What Odoo Actually Did
Odoo stopped Google Ads on February 1, 2026 and left them off until April 17. In January, the month before the shutdown, paid search cost €5.13M and produced 226,000 leads.
The hypothesis was simple. Cut the low-value leads, and let the sales team spend its time on the rest.
What a Paid Lead Was Worth
The deck grades paid leads against everything else Odoo gets, and paid loses on every line:
- They converted 3 to 7 times less than other leads.
- When they did convert, they were worth 25% less.
- Logo churn on paid customers ran 28.3%, against 18% for organic. MRR churn was 15.2% against 8.9%.
- Campaigns built on the word "Free" delivered 12% of paid ARR for 50% of paid spend.
The sales cycle came in at 37 days either way. So a paid lead does not close slower. It closes less often, for less money, and it leaves sooner.
Worth noting what Odoo does not do, because most companies its size do it without thinking: it does not bid on its own branded terms. Paying Google for the click of someone who already typed your company name is buying traffic you have earned.
What the Shutdown Cost, and What It Saved
Leads fell by 100,000 a month, about 20%, which is what Odoo expected. And there was "no organic lift from the shutdown." That line matters. The demand did not reroute itself through the organic door. It went away.
The cost: 1,000 to 1,500 subscriptions and an estimated €2M in new ARR. The saving: €10M in ad spend.
The Arithmetic, as I understand it
The deck reports enough inputs to work out the number that actually drove the decision.
€10M of removed spend cost about €2M of new ARR. That is €5 of advertising for every €1 of first-year revenue. The restart figures land in the same place from the other direction: €3M a month saved against €0.7M of new ARR lost, roughly 4.3 to 1.
Odoo's new rule is a payback target, which means €1 of spend has to bring back €1 of revenue inside a set window. North America's window is 12 months. Western Europe's is 6.
Measured against a 12-month payback, the spend Odoo cut was running about five times over the line. That is the whole story. It is not that advertising fails for Odoo. It is that the last few million a month of it were deeply underwater, and no amount of creative fixes a channel that is 5x off its own target.
When ads came back on April 17, they came back at 25% of the old budget: €1M a month instead of €4M. Cost per lead fell from €23 to €16, and volume dropped from 175,000 leads a month to 60,000.
Which sets the bar for the money Odoo kept. At €16 a lead, and €1,300 to €2,000 of ARR per subscription, about one lead in a hundred has to become a customer for that spend to pay back inside the year. That is the bet Odoo is making with the quarter it held onto.
(That math is mine, from the deck's figures. It treats revenue rather than gross margin, and it takes the €2M ARR estimate at face value.)
Two Things the Payback Target Leaves Open
First: payback on what? The Google invoice alone, or the invoice plus the sales time that works the lead? Those are very different fences. If a 12-month payback counts only media, the rep is free in the model, and the rep is not free.
Second: twelve months measures cash flow, not what a customer is worth. A payback rule rejects the customer who breaks even in month 13 and then stays for years. Odoo's own churn data says that customer is ordinary. A 28.3% logo churn rate implies an average paid customer life around three and a half years. Organic's 18% implies about five and a half. At €1,300 to €2,000 of ARR, a paid customer is worth roughly €4,700 to €7,000 in revenue before they leave.
Run the cut spend against that and it is still underwater, so lifetime value does not rescue the €3M a month. But it should move the fence. I asked Fabien in the comments how LTV figures into the analysis and whether it differs by channel, and when another commenter treated LTV as a fixed number, I pushed back. It is not fixed, and Odoo has better retention data than almost anyone in this market to prove it.
"Free" Was the Most Expensive Word in the Account
Every "Free" campaign is gone, because they bought 12% of paid ARR with 50% of paid spend. Taking the word out of the ad copy made the rest of the program more efficient.
Free attracts people who want free. That sounds obvious written down, and almost every B2B funnel I look at still leads with it. If yours does, pull that number this week. Odoo just published what it cost them.
What Fabien Says Is Still Broken
Two challenges on his own slide. The split between one-to-five-user deals and five-plus-user deals is hard to scale, and Odoo cannot generate five-plus-user leads at volume today. His candidate fix is value-based bidding, and he names the US as the market where it has to work.
Value-based bidding, defined. It is Google's family of bidding strategies that optimizes for the value of a conversion instead of the count of conversions. In Google's words, it "differs from conversion-based bidding (maximizing conversions within a given budget or using Target CPA), which aims to maximize conversion volume."
In practice, you stop telling Google "bring me more leads at €23" and start reporting what each conversion turned out to be worth, then run Maximize Conversion Value or Target ROAS so the auction chases the expensive lead that pays rather than the cheap one that does not. It needs at least two distinct conversion values flowing back, and in a B2B funnel that means importing offline conversions, ideally daily.
That is the lifetime-value argument in Google's vocabulary. Odoo already knows a five-user manufacturer is worth many times a one-user freelancer, and that paid churns at 28.3%. Feeding that back into the auction is what turns a blunt payback rule into an actual valuation.
Why This Hits the US Hardest
The appendix describes US Direct as "Dependant on CPC, low efficiency on paid" (their spelling). North America ran 53% of its leads off paid, more than any region except Mexico, and carries the 12-month payback target rather than Europe's 6.
So the market that leaned hardest on Google is the market where Google just got turned down by 75%, and nothing in the deck names the channel that replaces it here.
What I Would Do With the Recovered Money
Fabien has said Odoo events now draw 120,000 registrations a month, and that "sounds like we are also an event company now." That is the right instinct. It is not yet aimed at the US.
At Odoo Experience Americas in San Francisco I had between 90 and 100 conversations and met exactly four end users. One of the four came in through advertising. I also heard that Odoo reps are discouraged from bringing prospects to that show, which is a strange rule for an event company.
Here is where I would put the €3M a month, none of it on a billboard:
- Metro roadshows, with the leads handed to the local partner. Odoo already runs about 100s a year. Run them smaller, run them in more cities, invite more buyers, and give the attendee list to whoever serves that metro.
- Pop-up demo rooms. A week in one city, a storefront, walk-in demos by industry, a partner on site every day. Odoo demos better in person than any ERP I have watched. Let people walk in.
- Industry roundtables of eight to twelve buyers, hosted by a partner who knows that vertical and underwritten by Odoo. No booth, no keynote, no badge scanner.
- Open houses at reference customers. A plant tour at a manufacturer running Odoo beats every demo video ever made.
- The accountant channel. Intuit sourced 25% of new Enterprise Suite contracts through accountants. Odoo launched an official accountant network in Germany in August. Do that here.
Each of those produces a lead a partner can actually work, which is the only kind that clears a payback target.
What This Means If You Sell or Implement Odoo
Expect fewer handoffs from Odoo, and expect the ones you get to be better. The leads that disappeared are the leads that churned at 28.3%.
That is not a reason to wait. It is a reason to fix the parts of your pipeline that never depended on Odoo's ad budget: clear positioning, a real sales effort, and delivery that holds up after go-live.
The Bottom Line
Odoo proved with €10M of its own money that its marginal ad spend was five times underwater, and it was right to cut it. But the US is the market that leaned hardest on that spend, and value-based bidding is a fix for the auction, not a replacement channel.