Every shopper who's been doing this longer than a few months has a system. Might be a note in your phone. Might be labels typed onto addresses in Google Maps. Might be a spreadsheet you swore you'd keep up with.
Might be nothing, and you just remember the bad ones.
I did the Google Maps thing for about five months. Typed "NT" on maybe sixty addresses before admitting it wasn't working. Not because labeling is a bad idea. Because the label holds almost nothing, and I was doing it at the worst possible time.
What shoppers actually use
Four methods, roughly, and I've tried all of them.
Google Maps labels. You search the address, tap the label field, type something short. "NT." "Good tipper." "Cash $15." "DND." Then it sits there for next time you scroll past that spot.
A spreadsheet. Date, address, base pay, tip, sometimes a note. Usually Google Sheets so you can get at it from the truck.
Screenshots. Screenshot the offer, screenshot the tip notification, let them pile up in the camera roll.
Memory. Works fine for the customer who screamed at you about bananas. Not at all for the other four hundred stops.
Where Maps labels fall apart
A label is one string of text. That's the whole ceiling.
You can type "NT" but you can't type the history, because there's room for one answer and the customer has more than one order. Say they stiff you in March, tip $8 in May, tip nothing in June. What goes in the label? You overwrite it, or you cram "NT/8/NT" in there and hope you remember what the slashes meant four months later.
Amounts disappear too. "Good tipper" is not $6, and it isn't $22 either, and those are completely different orders when you're deciding whether to drive eleven miles.
You also have to be looking at the map to see it. The label doesn't come to you. It sits in Google Maps while the offer sits in the Shipt app with a timer running, so you're copying an address out of one app, pasting it into another, reading a two-letter note, switching back. On a bundle that's two addresses. Offer's usually gone.
The spreadsheet problem isn't the spreadsheet
Spreadsheets are great. A spreadsheet will hold every order, every tip, every address, and do math on all of it.
You just won't fill it in.
Fourteen stops on a Saturday. You get home at 9:40, you smell like the freezer aisle, and the tab is open on your laptop asking for fourteen rows. Nobody does that on day three. Some people manage it for a month.
The gaps are what kill it. A log with holes in it is worse than no log, because you'll look up an address, see nothing, and assume it's a new customer when they actually stiffed you twice in April and you just didn't feel like typing that night.
Screenshots aren't a system
They're a folder.
You can't search a screenshot for an address. Can't sort them, can't total them, and by month two there are six hundred images of offer screens in between pictures of your dog. The information is in there. Good luck getting it out.
The late tip breaks all of it
This is the part that makes manual tracking close to impossible, and it has nothing to do with discipline.
The tip doesn't land when you deliver. Sometimes it shows up an hour later. Sometimes the next morning.
So even if you're the rare shopper who logs everything right after drop-off, you logged the wrong number. You wrote down what you saw at the door. Fixing it means going back through yesterday's orders, matching each final tip to the right address, and correcting rows. Every day. Forever.
Nobody's doing that. I wasn't doing that. What you end up with is a log that quietly drifts away from reality, always in the same direction, making customers look cheaper than they are.
What actually has to get recorded
If you want tip history worth trusting, the address by itself doesn't do it. You need:
The address, matched consistently, so apartment 4B doesn't turn into three separate entries.
Every visit to it, not just the last one. Three orders at $0, $9, and $11 tell you something a single "good tipper" label never will.
Base pay sitting next to the tip. A $22 tip on a $9 base is a different customer than a $22 tip on a $40 base.
How long the tip took. Someone who always tips six hours later looks identical to a no-tipper at the door, and if you log at the door you'll write them off wrong.
Enough visits to see a pattern. One order is noise. Three is a customer.
If you're going to do it by hand
Some people will, so this is the version that holds up.
- One spreadsheet, not a mix of methods. Columns: date, address, store, base pay, tip, hours after delivery the tip landed, notes.
- Log the order at delivery with the tip column blank.
- Set an alarm for the next morning. Go into your Shipt order history, find yesterday's orders, fill in the final tips.
- Once a week, sort by address and look for repeats. Those repeats are the only reason you're doing any of this.
- Before accepting anything past about eight miles, check the sheet.
That's a real system. Ten minutes a day, and it works, as long as you do step 3 every single day. Which is where it always dies.
Why bother at all
Because the tip is most of the pay, and you're being asked to guess about it.
An order I ran last spring. 42 items, $16 base, customer 9.4 miles from the store. Shopping took 55 minutes with three substitution texts. Seventeen minutes out, five at the door, seventeen back toward the store. Call it an hour and thirty-five minutes, 18.8 miles round trip.
At 72.5 cents a mile, the IRS rate in effect that spring, that drive cost $13.63 in gas, tires, brakes, and resale value. Against $16 base, that leaves $2.37.
A buck fifty an hour, if the customer doesn't tip.
Same order with a $14 tip pays about $10.45 an hour. Same items, same drive, same 95 minutes. The only variable that moved was a number I couldn't see when I hit accept.
The offer screen gives you an estimate. An estimate is not what this customer did the last three times you shopped for them. And after a few hundred orders you've delivered to the same addresses over and over, so the information exists. It's scattered across a camera roll, a stale spreadsheet, sixty Maps labels, and your memory of a bad Tuesday.
Tip behavior repeats. That's the whole reason any of this pays off. Someone who tipped $12, $15, and $11 will land somewhere in that range again. Someone who's stiffed you four times isn't a mystery.
The trick is having it in front of you during the twenty seconds the offer is on your screen, instead of in an app you'd have to go dig through.
Auto Tip Map is an Android app that saves every order, tip, and address automatically, including tips that land hours after you've delivered, and shows an address's full history on the offer itself.