Two years of receipts, laid against what O&R's own riders said when their driver asked how they found GoRide. The gap between what the platforms claim and what customers confirm is the whole story.
Yelp's cost per customer went from $218 in 2024 to $539 in 2025 to $1,062 so far in 2026. Same monthly bill. Fewer customers each year. The price is climbing in Yelp's own numbers.
Every owner's first instinct is to question the source. Good. So let's name what we know, what we can't perfectly know, and why the answer still holds.
The customer counts in this brief come from the rider intake survey, where drivers ask paying customers how they found GoRide. It's the most honest tool the business has, because it's tied to a real, completed trip rather than a click on a screen.
It's also not perfect, and the report doesn't pretend otherwise. Drivers don't always ask. Some customers decline to answer. A meaningful share of trips are repeat customers, who would name "I've used you before" rather than a channel. So the survey numbers should be read as a floor, not an exact total.
A large share of riders are repeat customers, which means somebody acquired them in the first place. Across this two-year window, Google has been the dominant named first-time source every single year. The repeat base GoRide enjoys today was, in practical terms, built by Google. That doesn't reduce Google's value, it amplifies it.
Yelp's recorded customer count could be doubled and the directional finding would still hold. Yelp at $37,807 with 212 customers instead of 106 is $178 per customer, still 43 percent more than Google's $124 two-year average. The break-even point is roughly a 2.9x under-count, which would mean Yelp brought in 304 customers across two years that drivers never heard named. That's a stretch given the survey runs on every trip. The directional finding survives a doubling of Yelp's count, but it does not survive a tripling. The honest framing is that the gap is consistent, not catastrophic, and better measurement is still the right next move.
The Yelp app produces inquiries and conversations daily, which feels like the channel is working. Honest accounting separates that feeling from the result. Yelp's chat activity, even when it produces messages, is not producing booked trips at a rate that justifies $1,275 a month. It's a customer-service line the business is paying to operate, not a revenue engine.
Independent of the survey, Google's own data shows 7,731 visits sent to the site over 17 months and 960,304 appearances in search results. Most of those clicks, 73%, are people typing "GoRide" by name, meaning Google is delivering people who already know the brand and are ready to book. The survey can't see those people because by the time a driver asks, the customer just answers "I Googled you." Google's contribution is almost certainly larger than the named count.
In plain English: the data has room to be wrong, and the answer still doesn't change. That's why we can move on what's documented now, while building a better measurement system so the next finding is airtight.
24 months of spend against the customers who named each source on their trip. These are two-year averages, so the cost here is lower than the 2026-only figure above.
Same flat bill every month. Fewer customers every year. The price per customer keeps climbing.
Roughly half Yelp's spend, more customers, and the cost per customer holds about a third of Yelp's in every year measured.
The difference is timeframe. One averages across two years, the other is this year alone. Here's exactly what each one means.
Bottom line: $124 is the average across two years. $338 is this year alone. Either way, Google wins decisively. Yelp's matching numbers are $357 averaged and $1,062 this year.
Lower is better. Watch what Yelp does while Google holds the floor.
Spend, customers, and cost each, for both channels, every year.
Yelp's price per customer went up 5x while its customers fell from 70 to 6. Google's cost rose too, from $67 to $338, but it stays roughly a third of Yelp's price in every year. The $124 and $357 totals are these three years averaged together.
Google's own Search Console, the source it can't fake, over the last 17 months. If Google were dead, these numbers would be falling. They're climbing.
People are searching for GoRide and finding it. They are reaching the booking page. Visibility grew from about 30,000 monthly appearances early on to a peak of 117,000 in April 2026. That is a healthy, growing Google presence, not a broken one.
So when the Google Ads dashboard shows almost no bookings, that is not Google failing. It is the booking platform unable to report the sale back. The trips are happening. The scoreboard just can't see them. That distinction is the whole game.
Yelp counts every tap, every profile peek, every "get a quote" click as a lead. Most of those people never booked a ride. When O&R's drivers asked actual paying customers how they found GoRide, Yelp got named 106 times in two years. Not thousands. One hundred and six.
Plain terms, no jargon. Each one is fixable.
Roughly $1,275 a month, locked in, whether it brings O&R 6 customers or zero. The fee never moved. The results dropped every year. O&R is paying for a slot, not for customers.
Google is sending people and they are reaching the booking page, nearly 69,000 views of book-now in 17 months. The problem is that the booking system, bookridesonline.com, cannot accept a Google tracking tag. So the moment a rider books, Google goes blind. The booking happens. Google just never gets told. That is why the dashboard looks empty while real trips keep coming.
Google's display network and an older leftover campaign are running, but until conversion tracking is restored we can't say how much of that budget is genuinely producing trips versus producing untracked ones. The honest move is to put real measurement in place first, then audit by ad type and campaign, then optimize. We don't cut what we can't measure, and we don't justify what we can't prove.
An earlier read of Google billing pulled statement amounts, which net out credits and payment timing and run about a third below true ad spend. The corrected number, $20,512 over 24 months, comes from monthly net cost in the Google Ads console, which is the actual auction spend. Yelp's invoice screen separately shows duplicate charges that need to be reconciled. Before any platform number gets quoted to ownership, the dashboard claim has to be matched against what actually left the account.
Clear steps. Each one tied to money saved or made.
Pause Yelp ads for 60 to 90 days. Track two things: total booked trips, and what riders name on the survey. If trips hold steady, the chat engagement was activity without revenue and the savings stay permanent. If trips genuinely drop, Yelp goes back on with proof in hand. This isn't a guess, it's a controlled experiment that ends the debate either way.
The booking platform can't hold a Google tag, so stop waiting on it. Put a tracking phone number on the site to capture the calls Google sends, since a lot of airport and black-car trips book by phone. That alone makes invisible bookings countable, and it never touches the booking platform.
When a trip is confirmed in the system, feed that record back to Google as a completed sale so it links to the original click. This is the permanent fix and it works around the platform entirely. Set up once, accurate from then on.
Don't touch Google budgets until conversion tracking is restored. Once steps 02 and 03 are running, audit by campaign and ad type, drop what's truly not producing, and keep what is. This is a decision we don't have the data to make yet, but we will, soon. Patience here protects against cutting spend that's quietly working.
Take the freed Yelp and display budget and put it into Google search on LAX, airport, and shuttle terms. That's where high-intent riders already find O&R cheapest.
Reconcile both platforms against actual completed trips for three months. Then the Google-versus-Yelp call is settled with hard numbers, not screens.
Marketing is one line on the P&L. The two-year question, "where is the money going, why is no profit," deserves the same documented treatment on payroll, vehicle costs, fuel, insurance, dispatch, fees, and owner draws. This brief proves what a structured review can surface in one line item. The next 90 days should answer the rest of the question with the same rigor.
That's roughly what Yelp costs O&R per year for a handful of customers the business is likely already winning through Google. Redirected, it could buy multiples more riders.