中文简体When choosing the first battery swap location, the decisive variable is not rent, but whether the target riders will pass by on their route. The site is cheap, but if riders have to detour to get here, parking is inconvenient, and operations staff can't get in, then no matter how good the devices and batteries are, usage habits won't form. To judge whether a site is viable, first pass three conditions:Riders pass by frequently, stopping and swapping is convenient, and maintenance is accessible. Only when all three hold should you even discuss rent.
Why rent is not the main variable
Rent and location are not the sameClass variable: Rent is a predictable fixed expense; location determines the revenue ceiling.
The site fee is fixed monthly and has nothing to do with how many swaps happen today; swap revenue, however, occurs per swap. This difference is very intuitive at the negotiating table: cutting rent from 3,000 to 2,000 saves 1,000 a month, 12,000 a year—that's certain; but if the location is half a kilometer off, you lose a group of riders who would have passed by, and what you lose is revenue that keeps occurring every month, and it won't recover on its own.What you cut is a certain expense, but what you gamble is your entire source of revenue.
A more hidden layer: sites willing to lower their rent often already lack two things—rider density, and a reason to make people willing to stop.Landlords usually lower the rent not out of generosity, but because they know this location is hard to rent out.
How to judge the three conditions
First: whether target riders pass by often. It's "target riders," not "foot traffic." No matter how heavy the foot traffic in a commercial district is, if it's mainly pedestrians and private cars, it has nothing to do with battery swapping. Look at the routes of food delivery and courier riders, not the commercial district heat map.
Second: whether parking and swapping are convenient. Riders stay in front of the cabinet for only minutes. If they must climb steps, squeeze through narrow alleys, or compete with cars for parking spaces, the daily usable capacity of this location is capped by physical conditions.Convenience isn't shown by "whether you can swap," but by "whether you can queue up to swap at peak."
Third: whether subsequent maintenance is easy to reach. This one is the easiest to overlook. Operations need to regularly replenish batteries, clear faults, and do inspections. At sites that require access control, have traffic restrictions, or need detours, the time cost of every visit doubles, the average repair time gets stretched, and ultimately this feeds back into rider experience in the form of "this site always has problems."
Coverage radius follows the charge level, not the map
Riders do not swap batteries at "any location"; they only swap when their battery is nearly depleted.
An e-bike's single-charge range is on the order of a few dozen kilometers, while a rider travels far more than one charge's worth in a day, so they must recharge in between. The swap action happens when the remaining range drops near some psychological threshold—The more charge you have, the less worth it is to detour for a swap; the lower the charge, the less margin you have, and the smaller the range you can detour within.
So the optimal site appears at the intersection of the two constraints:By the time riders arrive here, their battery level is already low, and they do not need to take a detour.This is also why "placing it at the city's geographic center" often fails—the central location may, for each specific route, happen to be near the starting point where riders have the most charge. In practice, coverage is not a circle but the overlapping area of several lines: first draw the connecting paths between merchant clusters, then see where these lines intersect.
Prove it first, then scale: how big should the first batch be
The goal of the first site is not "rolling out a network," but "getting the service working within a small area"; the size of the first batch is determined by "whether it can be verified," not by "how many it can serve." Filling it up all at once based on "how many riders it can serve" has a problem: that denominator itself hasn't been verified yet. Multiplying it out to get a procurement quantity just magnifies the estimate one more time.The pace of expansion cannot exceed the pace at which operations and maintenance capacity is built.——Battery inventory gets diluted, the operations radius stretches longer, and usage at each site is too low to make the numbers work. These three things usually happen at the same time.
The costs at both ends are different: if the scale is too small, everyone queues at peak, and you can't tell whether the location is right but there aren't enough devices, or whether there are enough devices but the location is wrong—adding devices at this point will mask a site-selection error; if the scale is too large, redundancy will hide mistakes in the short term, and by the time it surfaces, the devices have already become sunk costs. The target is "just enough to fully run through one peak"—And a single cabinet can't produce valid conclusions, a single cabinet will almost certainly have queues during peak hours, and the cause cannot be identified; only with a second unit can you observe whether demand shifts away from the saturated one: if it shifts, demand is real; if both are empty, it is a location problem.
Don't fill batteries to match the number of slots.The number of slots is "how many batteries are charging at the same time," not "how many batteries can be served per day," In between lie charging time and battery-swap rhythm. The direction must be reversed: derive turnover from the measured number of peak battery swaps.Only with a measured rhythm can you set a ratio; copying "how many batteries per cabinet" only copies someone else's rhythm.
When to add investment—three criteria apply universally:Average daily turnover per cabinet, peak queue length, and the share of available batteries in the cabinet. But these three numbers have no industry-wide universal thresholds—city size, rider schedules, and site format all shift the baseline. So compare against your own last week and last month: if turnover rises and queues get longer, that's a demand signal; if turnover doesn't rise but idle batteries increase, that's over-provisioning.Comparing against your own baseline is more reliable than applying someone else's threshold.
Validation actions you can complete within a week
Site selection conclusions should not come only from "I think there are a lot of people here"—at minimum, there should be a week of verifiable records.
Action | How to do it | What to Look For |
|---|---|---|
On-site counting | At the same location, count 30 minutes each in the morning, midday, and evening | Number of two-wheelers passing by, and the proportion of riders among them |
Traffic-flow record | Ask 5–10 riders "where do you go from and to today, and where do you charge" | Whether their routes actually pass through here |
Stay and observe | Look at where riders actually park and dismount near this spot | Whether stopping to swap is smooth, and whether there are physical obstacles |
Site verification | Check incoming line capacity, ground hardening, drainage, and canopy conditions | Whether power and civil works upgrades require extra investment |
Operations site survey | Walk through it from the operations staff's perspective | How long it takes to get from the base to the site, and whether parking is easy |
These five items don't require complex tools, can be completed within a week, and cost far less than the price of one wrong choice.To judge whether a site is viable, there is only one standard: all three conditions hold at the same time. If even one is missing, don't use "cheap rent" or "let's try it first" to make up for it.