The Cost of Knocking 200 Doors: How Far You Actually Walk, Measured Across 1,200 U.S. Cities

Every guide to door-to-door selling quotes the same rule of thumb. Fifteen to twenty doors an hour in a normal neighborhood. Eight to twelve if it's rural. Twenty-five or more in city rowhouses. Fifty to seventy doors in a good day.

Those numbers are useful and roughly right. They are also entirely unmeasured. Nobody publishes where the gradient actually sits, because working it out means knowing how far apart American front doors really are — and that requires a geocoded address corpus rather than a survey of sales managers.

We have one. So we measured the ground.

Across 1,200 U.S. cities, the median city puts 31.8 metres of walking between one door and the next. Knocking 200 of them means covering 3.96 miles, about an hour and twenty minutes on foot, before a single conversation happens. In the most spread-out city we can name with confidence, the same 200 doors is 16 miles and five and a half hours of walking.

Mailboxes and Doorsteps Are Not the Same Unit

Before any of that means anything, there is a distinction the industry has never had to make, because until now no channel needed both halves of it.

Most of America sits near the floor. The median city measures 1.04 mailboxes per doorstep — one mailbox, one front door, no difference worth discussing. That is the country the rules of thumb were written for.

Then a quarter of it isn't. 25.7% of cities measure 1.5 or higher. At the ninetieth percentile the ratio is 2.0 — two mailboxes behind every entrance. The highest we measured is Aventura, Florida, at 9.19: a coastal high-rise market where 200 mailboxes sit behind roughly 22 physical entrances.

This is not a rounding detail. It is the point where the two channels stop being comparable at all. In Aventura, a mail wave of 200 reaches 200 addresses. A canvasser walking the same 200 mailboxes arrives at 22 lobby doors, most of them locked, and reaches nobody above the ground floor. Mail rides the elevator. The person knocking does not.

It cuts the other way too, and honestly: in the 702 cities we measured below 1.1 mailboxes per doorstep, the two units are effectively identical, and none of this changes the arithmetic at all.

How Far You Walk to Knock 200 Doors

Here is the distribution, measured across 1,200 cities. Each figure is the median over sampled anchor addresses in that city, walking to the 200 nearest doorsteps — apartments in one building collapsed to the single stop they actually are.

The first thing to notice is how narrow that is. From the tenth percentile to the ninetieth is a factor of 2.2. Most of the country is far more alike than the trade's rural-versus-suburban framing suggests. If you have knocked in two different mid-sized American cities, you have probably experienced close to the same walk.

The second thing is that the tail is where the money goes.

Tightest markets (converged cities only):

Widest markets:

Reading to Blairsville is a factor of 12.9. The same 200 doors, the same script, the same crew — and thirteen times the ground.

What That Actually Costs in Hours

Walking at a standard adult pace of about 1.3 metres per second, the median city's 200-door route is 1 hour 22 minutes of pure walking. Blairsville's is five and a half hours.

But the honest reading is not "walking eats your day," and it would be easy to overclaim here. In the median city, 31.8 metres between doors is about 24 seconds of walking per door. Set against a knock — approach, wait, the conversation or the non-conversation, the note in the CRM — walking is a minority of the clock. It is real, it is not the constraint.

In Blairsville, 129.3 metres is 99 seconds per door, and walking becomes co-equal with the knock itself. That is the actual finding, and it is more specific than the rule of thumb:

In most of America, distance is not what limits a canvasser. In the spread-out tail, it is roughly half the job.

Which is why the trade's published gradient — 15 to 20 doors an hour in the suburbs, 8 to 12 in rural territory — is a good description of the middle and an understatement of the edge. That ratio is about 1.7x. Our measured spread from the tenth to the ninetieth percentile is 2.2x, which lines up. The 12.9x between named extremes does not, because the rules of thumb were never built to describe the tail.

Why This Had to Be Measured, Not Modelled

The obvious objection is that none of this needed a corpus. Density is published for every city in the country; surely you can divide.

You cannot, and the reason is worth showing because it is the whole argument for measuring anything here.

If homes were spread evenly, 100 of them would sit inside a circle of radius √(100 / πD) at density D. That is straightforward geometry and it makes a clean prediction. Compared against what we actually measured, across 1,246 cities:

The model overshoots by roughly 75% at the median — and, more damaging, the error is not a constant you could correct for. It swings from 0.30 to 0.83 depending on the city. Homes cluster onto streets rather than filling space, and how tightly they cluster is itself a local fact.

Fit the relationship on logs and the slope against the predicted radius is 0.481, which works out to a density elasticity of −0.241 (R² 0.505). Our previously published radius report, computed independently and by a different route, put that elasticity at −0.24. Two derivations, same answer, which is a reassuring check on both.

The practical consequence: density does not give you distance, and the 100-door figure cannot be scaled by √2 to get 200. Every number on this page is sampled from the corpus for that reason.

Where Knocking Wins, and Where It Stops Making Sense

None of this is an argument that knocking is bad. It is the highest-conversion form of physical proximity marketing there is, and nothing else in the category puts a human being in front of a homeowner who just watched a crew work down the street. Published field-experiment research on door-to-door contact — decades of it, largely from civic canvassing — is the most rigorous evidence any proximity channel can point to.

What the measurement changes is the sizing question.

Knocking is labour-bound, and the labour scales with the ground. At trade-standard rates, 200 doors is a day and a half to two days of one rep's time in a typical market. In Blairsville it is meaningfully more, because five and a half hours of that is spent walking. Double the territory and you double the hours; there is no version where you don't.

Mail is postage-bound, and postage does not care about geography. 200 cards costs the same in Reading as in Blairsville. The thirteen-fold difference in ground simply does not appear on the invoice.

And in high-rise markets they are not even substitutes. At 9.19 mailboxes per doorstep, mail reaches recipients that no canvasser can physically get to.

The sequencing that follows is not complicated. Knock where the ground is tight and the doors are worth the hours. Mail where the ground is wide, where the buildings are stacked, or where you simply do not have the crew. And if you are going to knock a neighborhood anyway, there is a decent argument for warming it first — a job you already finished on that street, mailed to the homes around it, so the person answering the door has seen your name before it is attached to a stranger.

What We Measured, and What We Did Not

The route figure is a lower bound, deliberately. It is a greedy nearest-neighbour tour that moves from doorstep to doorstep in a straight line, ignoring streets, fences, culs-de-sac, dead ends and the backtracking every real route involves. A canvasser walks further than this — always. We publish the floor because the floor is defensible and the ceiling would be a guess.

Walking pace and per-door time are stated inputs, not our measurements. Where the trade's own benchmarks appear, they are reported as published.

And 112 of the 1,200 cities are excluded from the named rankings, because their measured median carried a confidence interval too wide to put a city's name against. They remain in the distribution statistics, where the uncertainty averages out. The ends of any ranking are exactly where noisy estimates collect, and a table of the most extreme cities is the easiest place in a report to publish an artifact by accident.

The spread within a single city is often as wide as the spread between cities — which is its own finding, and the practical version of it is this: your day depends on which street you start on more than it depends on which town you are in.