Rider management and delivery times: the alternatives that actually move your ranking

The minute you control is not the rider's, it's your grill's

When an order runs late, the courier is almost never at fault: the culprit is a badly declared prep time inside the aggregator. Assignment algorithms dispatch the rider based on when they calculate the food will be ready, so if your store declares 14 minutes and your kitchen actually takes 26, the driver lands twelve minutes early, waits, cancels, and the penalty falls on you. That steakhouse in Medellín was losing 41 orders a month to that gap, not to a shortage of motorcycles. It helps to size up the ecosystem where this gets decided: iFood reported more than 380,000 partner establishments across over 1,500 Brazilian cities in 2024 (iFood 2024), and Delivery Hero closed the same year with €48.8 billion in GMV, up 8% (Delivery Hero, Q4 and FY 2024 results). At that density of supply, the promised delivery window is the only ground where a small kitchen truly competes.

When does the aggregator's fleet stop being enough?

The aggregator's fleet stops serving you once you cross 60 daily orders with more than 70% of demand packed inside a 2.5-kilometer radius, because past that point the logistics commission you pay exceeds the cost of moving that same revenue yourself.

One number gives it away well before the P&L does: if your cancellation rate from rider waiting climbs above 4% for three straight weeks and your prep time is already calibrated, the problem stopped being yours. The second symptom belongs to demand, not operations. Deliveroo reported a record frequency of 3.5 orders per consumer per month across the UK and Ireland in 2024 (Deliveroo plc 2024); when your repeat base orders at that cadence and you still hold not one of their phone numbers, you are renting your own clientele month after month. Timing your ten best-selling dishes and setting the real prep time BY TIME BLOCK, never by average, costs nothing and is the first lever any operation must pull, no exceptions.

Option 1: surgical prep-time trimming, zero dollars

It takes two weeks: stopwatch in hand across five services, a different declared time for lunch and for dinner, and a rule that closes the store for twenty minutes whenever the kitchen queue passes six tickets. The steakhouse that dropped from 26 to 13 minutes of preparation regained 1.3 ranking positions in eleven days without touching its menu or its prices. Who is this for? For absolutely everyone, and most urgently for the owner about to sign a motorcycle contract. With a healthy food cost between 28% and 35% (National Restaurant Association), giving away margin on logistics before fixing the kitchen clock means paying twice for the same problem. Building an in-house fleet demands between USD 3,200 and USD 6,500 to launch and settles at roughly USD 1.80 per delivery in steady state, with a learning curve of ten to fourteen weeks before times stabilize.

Option 2: your own fleet with routing software

The profile that can carry it is narrow: operations from 60 daily orders up, a tight 2.5-kilometer radius, and someone who answers for the operation when it rains on a Friday. In return you recover margin and, above all, you recover customer data, the asset an aggregator never hands over. The downside is just as concrete: you inherit hiring, turnover, insurance, motorcycle maintenance and labor liability, a front that has cost Glovo, with over €1 billion in annual q-commerce revenue, litigation across several countries (EU-Startups 2025). If you don't want to manage twenty riders, don't. The hybrid settles the tug-of-war between margin and coverage by assigning peak hours to your own fleet and leaving the slow stretches to the aggregator, which is precisely where a fixed cost per motorcycle turns indefensible. If 62% of your orders land between 7:00 and 9:30 p.m., two owned bikes cover that block at close to USD 1.80 per delivery while the rest of the day travels on variable commission, with no idle payroll.

Option 3: a hybrid model split by time block

I got this wrong for years by recommending all-or-nothing: an owned fleet earns its keep BY THE HOUR, not by the month. The technical requirement is software that dispatches to both channels without duplicating tickets, since double-assigning one order wrecks more reputation than any delay. It fits operations between 40 and 80 daily orders with a sharp peak and a long trough. A ghost kitchen running three brands off one grill does not have three preparation times: it has a single one, wearing three costumes. A surge on the pizza brand poisons the bowl brand's times, and both slide down the ranking together, because the algorithm measures performance per store and grants no exemption for shared equipment. At Masterestaurant we call it cross-contamination of times, which is why Diego F. Parra advises that a dark kitchen starting from scratch launch with two brands cleanly split by shift rather than five fighting over the same fryer.

Cross-contamination of times in ghost kitchens

Industry structure data on U.S. ghost kitchens documents that concentration of operations per facility (IBISWorld, Ghost Kitchens US). Run the opposite test: separate the brands by shift for two weeks, and if times don't improve, then yes, your bottleneck sits in the equipment and not in the schedule. Last-mile automation is real, yet it is not mature for an operation under 200 daily orders, and committing capital there today means outrunning the payback by three years. Starship Technologies raised USD 90 million in February 2024 for autonomous delivery robots (Mordor Intelligence), and Wendy's rolled out its FreshAI system across 500 to 600 U.S. locations by the end of 2025 (CNBC 2024); those are chain figures, not independent-restaurant figures. Where a small operator does see immediate payback is in dispatch automation: rules that close the store on their own once the queue passes six tickets, alerts when real prep time drifts more than four minutes from what was declared, and a daily cancellation board.

Automation and robots: the option that isn't yours yet

That costs less than one motorcycle and pays for itself in the first month. If you run fewer than 40 orders a day, the honest answer is to stay on the aggregator's fleet and leave your logistics untouched. At that volume, two owned motorcycles load you with a monthly fixed cost no saved commission offsets, and the owner's hours drain into scheduling riders instead of into the menu, which is where your margin actually lives. Don't change either if your prep time still isn't timed by block: any alternative built on an intact bottleneck only makes the bottleneck more expensive. And when cash flow is tight —the leading cause of financial stress and closure among small businesses, according to Inc.— no logistics investment survives. Tomorrow, time your ten best-selling dishes with a stopwatch and compare each one against the time your store declares today. ALTERNATIVE 1 — Surgical prep time reduction inside the aggregator.

The five alternatives, without romance

Cost: 0 USD. Curve: 2 weeks. For whom: everyone, before evaluating anything else. It means timing your ten best sellers, setting real prep time by daypart instead of a daily average, and pausing the store for 20 minutes whenever the kitchen queue passes six tickets. One grill house that went from 26 to 13 minutes recovered 1.3 ranking positions in eleven days, with no menu or price change. ALTERNATIVE 2 — In-house fleet with routing software. Cost: 3,200-6,500 USD upfront and 1.80 USD per delivery at scale. Curve: 10-14 weeks. For whom: operations from 60 daily orders with 70% of demand inside 2.5 km. It buys margin and returns the customer record, yet it hands you two problems that were never yours before: accidents and turnover. Budget 40% annual turnover and a part-time coordinator from day one. ALTERNATIVE 3 — Hybrid fleet by daypart.

The five alternatives, without romance — in practice

Cost: 60% of a full fleet. Curve: 6 weeks. For whom: restaurants with two sharp peaks and a long valley. Your bikes cover 12:00 to 14:30 and 19:00 to 21:30, the aggregator absorbs everything else including late nights. This is the only alternative I have seen hold margin and coverage at once in mid-size cities, and also the one that demands the most discipline on the board. ALTERNATIVE 4 — White-label third-party logistics. Cost: 2.10-2.60 USD per delivery, no assets, no payroll. Curve: 3 weeks. For whom: a virtual brand or ghost kitchen selling through its own channel that wants no bikes on the balance sheet. You keep the customer and the WhatsApp conversation, pay per executed delivery and avoid dead capital. The limit is real: many secondary cities have no serious operator and coverage collapses on Sundays. ALTERNATIVE 5 — Pickup plus refrigerated locker with an aggressive incentive.

The five alternatives, without romance — key points

Cost: 900-1,800 USD for the unit, 0 USD per order. Curve: 4 weeks. For whom: kitchens with offices, a campus or a gym within 600 metres. A 12% discount for collecting converts 18-24% of nearby orders, and that share comes straight out of logistics cost. It is the most ignored alternative in local foodtech and the best return per dollar invested.