Why "$1 a mile" isn't your rule
The folk rule floating around every driver forum is to take it if it pays $1 (or $2) a mile. That has a real problem: it prices every driver's time and car identically, and it prices a 2-mile order the same as a 12-mile one. Two drivers sitting in the same restaurant parking lot can have genuinely different costs per mile (different fuel economy, different pump price that week, different repairs waiting down the road) and genuinely different ideas of what an hour of their evening is worth. And every delivery carries a fixed chunk of time that a per-mile rule can't see: the wait at the restaurant, the walk to the door. Small offers live or die on that fixed time, not the mileage.
The honest math has three parts: what the miles cost your car, what the minutes are worth against your target, and what the offer pays. Put together: an offer clears your target when it pays more than your stop-time cost plus your true per-mile cost times the miles. That's the whole site in one sentence. The calculator above does it live, and the Decline Sheet does it in advance.
Worked example
Miles you drive but don't get paid for
A 7-mile delivery that ends next to a busy restaurant strip and a 7-mile delivery that strands you at the edge of the map are not the same offer. This calculator handles that honestly: a drop-off in the sticks doesn't make the route cheaper; it adds reposition miles (and their time) to the cost side. Miles you actually drive always count; a good destination is valuable because it makes the next offer come sooner, not because this one becomes free.
Where the verdict comes from, line by line
There is no black box in this tool, and it's worth watching the arithmetic run once so that you trust it later. Everything below uses the starter numbers this page suggests: thirty cents a mile of vehicle cost, an $18 hourly target, three minutes to cover a mile, five minutes of standing around per order.
Say a $10.00 offer comes in for 7.0 miles. Start with the clock: five minutes at the counter, plus three minutes for each of the seven miles, is 26 minutes of your evening. Then the car: seven miles at thirty cents is $2.10 of fuel, tread and wear. What survives is $7.90, and $7.90 earned across 26 minutes works out to $18.24 an hour. That clears the $18 target, so the panel turns green. Money on this site always rounds up to the nearest cent, which is why the display reads $18.24 rather than trailing off into decimals.
The Decline Sheet reaches the same place from the opposite direction, and this is the half worth memorizing. Every delivery bills you before it bills you a single mile: five minutes of an $18 hour is $1.50, and you pay that whether the drop-off is around the block or across the county. After that, each mile charges you twice: thirty cents of vehicle cost, and three minutes of your hour, which is another ninety cents. A mile therefore costs $1.20 all in, and the rule prints itself: $1.50 for the stop, $1.20 for every mile. Seven miles means $1.50 + $8.40 = $9.90 of ground to make up, the offer pays $10.00, and you are clear by a dime. Same verdict, no typing, no phone.
Why the sheet always rounds down
Break-even mileages rarely land on tidy numbers. Under these same settings a $7.00 offer breaks even a little past 4.58 miles, and the sheet prints 4.5, not 4.6. That is deliberate. Driven exactly 4.5 miles, the offer returns $18.33 an hour and you are safely above your target; driven 4.6, it returns $17.94 and you have quietly missed. A cheat sheet you have to second-guess is worse than no cheat sheet, so every figure is floored. The number you see always clears your target when you drive precisely that far.
Dollars per mile ranks offers; it can't judge them
Per-mile is a decent sorting tool when two offers are otherwise alike, which is why the verdict line still shows it to you. It falls apart the moment the offers differ in length, because it cannot see the fixed cost of the stop.
Watch two offers under the same starter profile. Offer A pays $10.00 for 7.0 miles, which is $1.43 a mile. Offer B pays $6.00 for 4.0 miles, which is $1.50 a mile. On per-mile alone, B wins and it isn't close. Run the clock instead: A takes 26 minutes and returns $18.24 an hour, while B takes 17 minutes and returns $16.95, under target and down in the close-call band. The shorter order looked richer per mile and paid worse per hour, because its unavoidable five minutes at the restaurant were spread across four miles instead of seven.
The general shape of it is this: a per-mile rule flatters short orders and penalizes long ones, which is precisely backwards from the way delivery time actually accumulates. Any rule that ignores minutes will hand you that same error over and over, all night, and it will feel like bad luck rather than bad arithmetic.
What the trip back actually costs
Take a $7.50 offer running 5.0 miles. If it drops you somewhere the next ping arrives quickly, so no repositioning is needed, the math lands on exactly $18.00 an hour — dead on target. Now switch the control above to deadhead back. The same offer now carries 10 cost miles and 35 minutes, its vehicle cost doubles to $3.00, and the estimate falls to $7.72 an hour. The payout did not change. The only thing that changed is where the delivery leaves you standing when it's over.
That gap is the entire argument for taking the return trip seriously, and it's also why the tool refuses to let a convenient destination make an offer look cheaper. A drop-off beside a busy restaurant strip is genuinely valuable, but the value arrives as the next offer coming sooner, not as this one costing less to run. Miles you actually turn the wheels for always land on the cost side. Answer the reposition question honestly and the calculator will tell you something the offer screen never will: that a few of the best-looking payouts on the board are the ones that strand you.
Working the sheet on a real shift
The whole point of the Decline Sheet is that you never do arithmetic with a phone in one hand and an offer counting down in the other. DoorDash's driver help pages put that window at 50 seconds. Fifty seconds is ample time to read one line and act. It is nowhere near enough to model a delivery.
So the routine goes: fill in the profile once, glance at the two-line rule on the red card, then print it or leave the tab open. In the moment you are running a comparison, not a calculation. The offer names a dollar figure and a mileage figure. Your card names what the stop is worth and what each mile is worth. If the offer beats the card, it clears. If it's close, it's close, and the texture of the night breaks the tie: how long you have been sitting idle, how far the drop-off drags you from the busy zone, whether you want to be home by ten.
The table underneath does the same job for anyone who would rather scan than multiply. Common offer amounts run down one side; beside each is the furthest you can drive and still hit your target. Seen enough times, those pairs stop being a lookup and start being a reflex. If you also filled in a bare-minimum hourly, a second column appears marking your absolute walk-away line, which earns its keep late in a slow shift when the honest question is no longer "is this good" but "is this better than sitting here."
Print it. A sheet of paper on the passenger seat cannot ring, buffer, or lock itself.
Train the gut, then trust it
Use the live calculator between offers and after the shift rather than during one. Punch in the offer you just declined and find out whether you were right. The history list keeps your last ten, so you can watch your instincts calibrate against the arithmetic instead of against a feeling. The goal is to stop needing either tool for the obvious offers and to keep it for the ones sitting near the line.
What this calculator cannot see
An honest tool should say where it's blind. This one knows a payout, a distance, and your own pace assumptions, and that is the sum of its knowledge. It has no idea about the drawbridge, the stadium letting out, or the fact that this particular kitchen runs twenty minutes behind on Fridays. It cannot tell a townhouse with a driveway from a fourth-floor walk-up in a building whose elevator has been broken since spring. It does not know about downtown parking, the gated community with a callbox nobody answers, or the apartment complex numbered by a scheme known only to its architect.
It also has no way of knowing whether this order gets stacked with another one halfway through, which rewrites the arithmetic entirely and rarely in your favor. It cannot predict what the weather does to your pace or to your brakes. It holds no opinion whatsoever about the neighborhood.
Every one of those blind spots lands in the same place: your minutes. Each of them is really a claim that this particular delivery will take longer than three minutes a mile plus five at the stop. So when you look at an offer and something in you mutters "that one's a slog," you are making a pace estimate, and you are probably right. The verdict is a baseline to argue with, not a ruling to obey. Decline anything you don't want to drive.
The pace fields are the honest fix for the recurring version of this problem. If your market genuinely runs slower than three minutes a mile, or your usual pickups genuinely hold you longer than five minutes, change those two numbers and everything on the page (verdict, rule, table) moves with them. Time yourself across one mile and one pickup on your next shift. It is the highest-value ten minutes you will spend on this site.
About the number in the offer box
Type in whatever total the offer screen shows you, because that is the figure you are actually deciding on. Just be clear with yourself about what it is. Where a platform displays an estimated total that includes a tip before the delivery has happened, that estimate is a forecast of something the customer has not finalized, and it can still move after you have already done the work. Only two apps publish a window for a customer moving it down: Instacart, where a tip cannot be decreased after two hours, and Spark, where customers get three hours to adjust one. No other platform publishes a comparable downward window.
The sensible way to hold this: treat the verdict as a judgment on the portion you can count on, and treat anything above it as upside. An offer that only clears your target if the tip behaves exactly as advertised is not really clearing your target. This site deliberately publishes no tables of what a platform pays or what tips typically run, because those figures move by market, by hour and by week, and a confident average from a stranger on the internet is worth less to you than one week of your own records. The platform pay page covers how the pay structures are assembled; what they produce in your zip code is a question only your own history can answer.
Which is the argument for keeping records at all. What you really made takes a finished shift and works the same math backwards.
Where the numbers on this site come from
The 50-second decision window is DoorDash's own figure, published on its driver help pages. The vehicle-cost side is assembled from things you can verify yourself: pump price measured against your actual fuel economy, plus maintenance and wear. For a placeholder before you have receipts of your own, AAA's 2025 Your Driving Costs puts maintenance, repair and tires at 11.04 cents a mile in its new-vehicle averages. The cost per mile page walks through building the figure properly, and whatever you land on there flows into this one.
Two things are worth knowing about the tolls and parking box, both from IRS guidance. Business parking fees and tolls are deductible in addition to the standard mileage rate, and records kept at or near the time of use carry more weight than a reconstruction attempted the following April. So the dollar you enter tonight is not merely a cost charged against this offer; it is a note to your future self. Commuting miles are generally not deductible, which is its own reason to be precise about when a shift begins. The mileage deduction page handles the rates and the split-year situation. All of that is federal and general, and none of it is advice about your particular return.
The Decline Sheet formula was derived, then independently re-derived, before it went anywhere near this page, and an automated harness re-checks it on every build across hundreds of randomized profiles to confirm that a mileage the sheet prints always meets the target it promises. If you ever catch a case where it doesn't, the feedback link at the bottom reaches a person.
Frequently asked questions
How do I know if a DoorDash order is worth taking? Compare the offer against your own numbers, not a folk rule: subtract your vehicle cost per mile (gas, maintenance, wear) from the payout, divide by the realistic minutes the delivery will take, and check the resulting hourly rate against your target. DoorDash gives you 50 seconds to decide, so the practical approach is a pre-computed decline sheet: for each offer amount, the maximum miles at which it still meets your target.
What is a decline sheet? A personal cheat sheet computed from your cost per mile, target hourly rate, and typical delivery pace. It lists, for each offer amount, the most miles the delivery can be and still meet your target. For example, at $0.30/mile costs, an $18/hour target, and a typical 3 minutes per mile plus 5 minutes per order, a $7.50 offer is worth it up to about 5.0 miles. It collapses to one memorizable rule of the form 'take it only if it pays at least $1.50 for the stop plus $1.20 for every mile.'