Eduardo Ramirez drives for Uber in Houston, Texas, six days a week, mostly evenings and weekends. He is twenty-nine years old, drives a 2020 Toyota Camry hybrid he bought used at the start of 2024, and treats the rideshare income as the primary source of household money for himself and his partner. The Camry hybrid is rated at 52 mpg combined by Toyota. Eduardo had been operating, before he started logging, on the assumption that he was getting roughly that figure on his routes. He had no specific reason to think he was wrong.

On February 1, 2026, the day after the U.S.-Israel strike on Iran and the day the gasoline futures market opened sharply higher, Eduardo started keeping a fuel log. The log was not part of any plan to optimize his driving. It was a defensive reaction to the news. He wanted to see, week by week, how much the change at the pump was costing him. He decided to do it for ninety days and then look at what the data showed.

It is now early May. Eduardo has ninety days of data on the Camry. He filled the tank approximately every other day across the three months — the rideshare driving pattern compresses a year of normal driving into about four months of high-frequency fills. He has 47 separate fill records. Each fill captures the date, the time, the station, the price per gallon, the gallons pumped, the odometer, and the trip-meter mileage since the last fill. The total dataset is about 4,200 miles of driving, $1,470 of fuel, and 47 individual transactions across the four major Houston-area gas-station brands he passes on his routes.

The data has shown him three things he had been wrong about, two things the rideshare app was specifically costing him in fuel, and one driving-habit change that paid for itself by week three. None of the discoveries required Eduardo to know anything about cars beyond what he already knew. They required him to write the four numbers down at every fill and then look at the dataset.

What he had been wrong about, number one — his actual fuel economy

Toyota's combined figure for the Camry hybrid is 52 mpg. Eduardo's actual combined figure across the 47 fills, on his actual driving, is 43.8 mpg. The gap is approximately 16% below the manufacturer's stated figure — significantly worse than the 5-9% gap most non-hybrid vehicles run versus their EPA combined number.

The gap has a specific cause. Hybrid vehicles get their advertised combined figure under driving conditions that include a mix of low-speed urban work (where the electric motor does most of the propulsion and the engine cycles off) and steady highway cruising. Rideshare driving is dominated by stop-and-start traffic on surface streets, idling at pickup locations, and short bursts of acceleration as the driver weaves toward a fare's coordinates. The Camry hybrid, in this kind of work, runs the engine far more than its EPA combined figure assumes. The result is the 16% gap.

For Eduardo, the practical consequence is that his fuel cost per mile is not $0.078 as he had assumed (4.06 / 52). It is $0.093 (4.06 / 43.8). On a typical 220-mile night of driving, the gap is about $3.30 — money he had been pocketing in his head and was actually spending in the tank. Across a six-day week, the gap is roughly $19.80. Across a month, it is $86. Eduardo had been silently overcounting his nightly take-home by about $86 a month.

The data did not change the underlying economics of his driving, but it changed his planning. He raised his minimum acceptable fare price (the floor below which he declines a ride) by 12 cents per mile, which the Uber app permits drivers to do indirectly by being more selective about which pings they accept. He stopped accepting cross-city pings to the airport during peak gas-pump periods because the deadhead-back-from-airport leg is where the Camry's hybrid system is at its least efficient. The data made the floor visible.

What he had been wrong about, number two — the cheapest gas station on his route

Eduardo had a default station — a Buc-ee's near his home in northeast Houston where the regular unleaded was always among the cheapest in the area and the convenience store was a pleasant place to take a five-minute break between fares. He had assumed, without checking, that the Buc-ee's was his cheapest option.

The 47-fill dataset showed it was not. The cheapest station on his actual routes, averaged across the ninety days, was a no-name independent on Wayside Drive in southeast Houston that he passed approximately three nights a week on the way back from suburban fares. The independent's price was on average 17 cents per gallon less than the Buc-ee's across the dataset. The savings on a 12-gallon fill was $2.04. Across the 22 fills he could plausibly have shifted from the Buc-ee's to the independent, the savings would have been about $45 over the ninety days. Annualized, the savings would be approximately $180 — for the cost of choosing a different station on nights when his route already passed it.

The discovery was not earth-shattering, but it illustrated something Eduardo had not appreciated: the variance in pump price across nominal-quality unbranded stations and major-brand stations on the same Houston roads is consistently 12-20 cents per gallon. The variance is invisible to a driver who has a default station. It is visible to a driver who has a log.

What he had been wrong about, number three — when the price actually moved

Eduardo had assumed, with most American drivers, that pump prices moved on a roughly weekly cycle and that there was not much point in trying to time fills against the cycle. The 47-fill dataset showed the actual pattern in his market is more granular than that. Across the 90 days, the cheapest fills (relative to the rolling 7-day average price) clustered between Tuesday afternoon and Thursday morning. The most expensive fills clustered between Friday evening and Sunday afternoon. The gap between the cheapest day-of-week and the most expensive day-of-week, averaged across his fills, was approximately 9 cents per gallon. Not enormous. But consistent.

Combined with the station-choice savings, the day-of-week adjustment cuts another $15 a month off Eduardo's fuel cost. He now refuels, when he can, on Tuesday or Wednesday afternoon rather than Friday night. The change cost him zero. The savings annualize to about $180.

What the rideshare app was costing him, specifically

Two patterns emerged from the data that Eduardo had not expected to find at all. The first was a fuel-economy penalty associated with a specific type of fare. Pings that took him to addresses inside large apartment complexes — where the pickup location is somewhere within a maze of parking lots and the rider often takes 90-180 seconds to find the car — produced fuel-economy on the trip-back that was roughly 8% below his average. The cause is the idling at pickup. The Camry hybrid's engine cycles on at idle when the climate-control system requires sustained AC, which in Houston between April and October is most of the time. Eduardo started declining apartment-complex pings during peak summer hours. The fuel savings was modest, but the time savings was real.

The second pattern was bigger. The Uber app's "destination filter" feature — which lets drivers indicate they want rides heading toward a specified destination — tended, in his data, to route him onto secondary routes that produced fuel economy 12-15% worse than his average. The reason is that the algorithm prioritizes pings that bring him geographically closer to his stated destination, which often means accepting fares that detour onto stop-light-heavy surface streets rather than highway routes. Eduardo turned off the destination filter except for true end-of-night runs and his fuel economy on his last two hours of driving improved by about 2 mpg. At his fill frequency, that is approximately $11 a week saved.

The driving-habit change that paid for itself by week three

The single largest fuel-economy gain in Eduardo's ninety days came from changing one habit: how he accelerates from a stop. The Camry hybrid is engineered such that gentle acceleration (specifically, accelerator pedal travel of less than about 30%) keeps the gasoline engine off and uses only the electric motor for the first 8-15 mph of speed. Aggressive acceleration above the 30% pedal threshold engages the engine immediately. Eduardo, like most Houston drivers, had a habit of accelerating briskly off the line in traffic. The data showed that fills following nights with longer routes through stop-light corridors (where he was repeatedly accelerating from stops) returned 5-7% worse fuel economy than fills following highway-heavy nights.

The fix took about a week of conscious effort. Eduardo began counting "one Mississippi, two Mississippi" between the green light and his right foot reaching the accelerator. The Camry crept forward on its electric motor. By the count of three Mississippi, the engine was already starting to engage and Eduardo could press the pedal with the engine running rather than starting it. The result was that his time to 25 mph was about a second slower than before. His fuel economy across the next two weeks improved by 2.4 mpg. At his driving frequency that is approximately $94 a month saved. Across the eighteen months the EIA expects elevated fuel prices to persist, the same habit, sustained, is worth approximately $1,700.

The dataset is the discovery

None of the three "wrong about" findings, two app-cost findings, or one driving-habit finding were available to Eduardo before he started logging. They were not findable through reading articles, asking other drivers, or guessing. They came out of his own data on his own driving and were specific to his vehicle, his city, and his routes. A driver in Atlanta with a Prius would find different patterns. A driver in Sacramento with a non-hybrid Civic would find different patterns. The discoveries are not transferable. The method — log every fill, write four numbers, look at the data after ninety days — is.

The total time investment was approximately five minutes a week, spread across the fills themselves and one weekly five-minute session reviewing the dataset. The total return, annualized, is approximately $2,400 across the savings findings combined — not including the downstream effects (reduced engine wear from gentler acceleration, reduced brake wear from the same, slower depreciation from a vehicle that is being driven within its design envelope).

For a rideshare driver operating on margins that have already been compressed by the rise in fuel costs, $2,400 a year is not a luxury. It is approximately three weeks of net income recovered from data Eduardo already had access to but had not been writing down.

What the same exercise looks like for a non-rideshare driver

Most American drivers will not see findings of the magnitude Eduardo saw, because most American drivers do not put 4,200 miles on a vehicle in ninety days. The compressed-time effect of rideshare driving makes patterns visible faster. For a typical commuter driving 12,000 miles a year, the same ninety-day log will show smaller, slower-emerging versions of the same kinds of patterns: the cost-per-mile baseline, the day-of-week price variance, the station-choice variance, the driving-habit penalty, the fuel-economy drift that signals a deferred service.

The findings, scaled to a normal driving pattern, typically work out to $200-$600 a year in savings for drivers who do something with the data and approximately zero for drivers who collect it without acting on it. The dataset is the discovery; the action is the value.

The My Mekavo fuel log, free, in your pocket

The My Mekavo driver portal includes the fuel-tracking tool Eduardo would have used if he had not built his own spreadsheet. It captures the four numbers per fill, computes the cost-per-mile, fuel-economy, and trend graphs automatically, and lets you compare costs across stations and across vehicles. It is free. It does not lock the data in — you can export it to a spreadsheet at any time. It also lets you, optionally, attach a receipt photo to each fill (useful for self-employed drivers tracking deductible business mileage).

The point of the tool is the point of Eduardo's ninety days: when fuel costs are this volatile and this elevated for this long, the difference between a driver who can see what their vehicle is actually costing them and a driver who cannot is roughly an order of magnitude larger than it has been at any time since the early 1980s. The tool exists because we wanted to make the visibility free for every American driver who wants it.

Where Eduardo's log goes from here

It is May. Eduardo has not stopped logging. The Camry will get its next service in about three weeks; the log will tell him whether the service restored fuel economy that has begun to drift below his April baseline. His partner has started a log for the household's second vehicle. His brother, who drives a Ford Transit van for a delivery contract, asked him for the spreadsheet template. The log has spread within his household and his family.

The energy environment that pushed Eduardo to start logging on February 1 has not changed. The EIA still does not see retail prices below $3 before 2027. What has changed, in Eduardo's household, is that the family knows what its vehicles cost. That knowledge is the smallest unit of control available to an American driver in 2026, and it costs five minutes a week to acquire.

Official sources cited in this article

Last updated: April 2026. Individual driver findings depend on vehicle, region, and driving pattern. The dataset described in this article is illustrative; specific savings figures will vary.