Vending machine business plan

By Tom Hill, Editor. Updated November 12, 2026.

Locations come first, and they decide the plan

A vending plan without signed locations is a guess, because the host decides whether you stay, sets the commission and can unplug your machine¹¹. Tom Hill, who edits this site, says, “location is the most important thing, and the business will fail if you can’t get a good location.”

Sales are lopsided. On one Utah route offered for sale, the top 3 locations made 76% of its $31,000 a year. The best site did $415 per machine a month; the weak tail about $101⁴.

The rule owners repeat most is to get the location’s yes before you buy the machine⁷. Their tests for a site: 25 to 40 employees⁶, or, says one new operator, 50 or more⁸; skip anywhere with fewer than 10 cars in the lot¹⁰; and “blue collar accounts are always better than white collar”¹⁰.

Expect the hunt to take a while. One ex-owner counted 20 no’s for every yes¹¹, though another got 6 yeses on his first day of door-knocking⁹. The best sites are often taken: the four biggest firms hold 46.4% of employer vending sales¹⁶.

Write down whether commission is a share of gross or net, and whether sales tax comes out first. And plan each site by season: a campus route fell to about 15% of its term-time sales over summer¹².

Machines and how to pay for them

Plan on a card reader: 78% of sales on Cantaloupe’s devices were cashless in 2025²⁵.

These paybacks leave out delivery and first stock. The $136 is 5 self-reported accounts²; the $50 is from the P&L below.

Borrowing stretches those paybacks further. A 7-year ex-operator’s warning: “Tons of people lose at this game because they get into debt.”⁶

Turnkey packages such as Naturals2Go worked out at $9,666 to $11,999 per machine, about 1.5 to 3 times the new-machine medians above, and “All Sales are final”³². Those are package prices, so check what the disclosure says each package includes before you compare.

The per-machine P&L

This worked example shows one machine’s month at three kinds of site, built from sourced costs. It doesn’t predict what your machines will make.

Lines are rounded to the dollar; the bottom line is worked from the unrounded figures, so columns may not add exactly.

Our $50 sits well below the $136 owners report (5 self-reported accounts)², because two of those five pay no commission, their figures don’t say whether storage, gas or repairs were counted, and those who post numbers are mostly still in business.

Build these into your own version:

• Product cost depends on what you stock. At public vend prices, store-brand water costs about 4.3% of its price, Pop-Tarts 20.8%, a 12 oz can 44.5% at $1.25, and a Snickers about 56% at $2.00²⁷. The “double your cost” rule fails for full-size chocolate.
• Card fees hit quiet machines hardest. The monthly fee doesn’t shrink with sales. At 100 card sales a month, fees take 12.6% of a $1.50 sale on Cantaloupe Engage and 19.2% on Cantaloupe One²³.

• Sales tax comes out of the price. Texas taxes 100% of soft-drink and candy receipts at 6.25% to 8.25%²⁹; California taxes cold food on 33% of receipts²⁸. The table doesn’t deduct it.
• Allow for a slow start. Stock takes cash before the machines earn it back: one test spent about $2,700 on first stock for 4 machines¹⁴, and Harber reckons building up stock takes about 3 months of a route’s net⁴. One ex-owner says break-even took “nearly a year… if everything goes right”¹¹.

How many machines for $100k (or for $1,000 a month)

The $363 is one operator described second-hand on a podcast, not checked¹⁵.

The often-quoted $6,284 per machine a year is a 2023 forecast averaged across operators with about 277 machines each²⁶, so don’t plan a new route on it.

Time is the other limit. One ex-operator puts one person’s limit at about 40 machines a week, if they’re close together⁹, so 62 machines means a van, storage and probably help. From 20 machines, FDA calorie labelling applies³⁹.

A realistic first route is 10 machines at the median, which makes a side income: about $2,670 a month in sales, and about $500 (our costing) to $1,360 (the owner median) left over².

Buy a route or build one

Buying a route skips the location hunt, but you pay for the seller’s numbers.

Value a route on what routes sold for, not on what sellers ask. In BizBuySell’s sold data the median route went for 93% of its asking price³⁴. Two asking prices we read sit above the 1.11× sold median: $60,000 asked for about $45,000 a year in sales (1.33×)³, and the Utah route at $40,000 for $31,000 (1.29×; its write-up says 1.9×)⁴. One buyer paid $30,000 for a hotel route claimed at $3,000 to $5,000 a month that really made about $2,000³.

For a route purchase, the plan is mostly checking the seller’s figures, then running the P&L above on their real sales.

Who will read your plan, and what it costs to write

• Just you. The SBA says a lean plan is “typically only one page”³¹. The P&L and site list above are most of it.
• A bank or SBA lender. See the lender box below the price table.
• A franchisor or “business opportunity” seller. Their disclosure sets the numbers. Canteen only franchises people already running vending or food service ($9,150 to $44,300 to start)³³. Naturals2Go is sold as a business opportunity and makes no earnings claims³².

The route costs are fixed, which is why 5 machines leave so little. Copy the layout and use your own sites.

Tom’s view as editor is that vending “needs a clear business plan”. If the plan is for you, our ranking:

1. Free: the SBA lean format plus ProjectionHub’s free vending projection template⁴⁴, if you’re comfortable in a spreadsheet.
2. Plan software, median $17/month. It’s the next cheapest route and gives you structure if you’d rather work through a template on your own.
3. Draper’s Biz Plan workshop, US$25/month. We make it. It’s a guided workshop rather than a blank page, which suits someone writing a first plan, and a month of it costs a few dollars more than a month of LivePlan and less than half the one-off median paid template. Pick it over plan software if you want to be walked through the questions rather than fill in a template. We haven’t checked whether it keeps separate weak, median and strong P&Ls per site, so keep the table above alongside it.
4. A paid vending template. Its example numbers aren’t yours, so it doesn’t test your sites.

What can go wrong

1. The site is weak, lost or seasonal. A hotel with its own snack shop³; a staff lounge too quiet to pay back¹³; a host whose roof collapsed⁶; on a bought route, a host who may unplug your machine if nobody introduced you¹¹.
2. You overpay. Debt for new machines, or a route bought on inflated figures³.
3. Repairs and expired stock. Boards and readers at $300 each⁴. One commenter, not an operator, says a hired repair “will usually cost your entire months worth of profits”⁶. Stock expired even at 10 machines¹¹.
4. Locator and turnkey scams. The FTC’s Business Opportunity Rule covers sellers who promise locations: they must give you a one-page disclosure at least 7 days before you sign or pay, and back any earnings claim in writing³⁶. In 1998 the FTC and states brought 40 actions in “Operation Vend Up Broke”³⁷. In 2008 a court entered a $4.06 million judgment against Westbrook Marketing³⁸. Today it shows up as locator spam in Reddit threads⁸.
5. Permits vary. Wisconsin charges $125 plus $9 a machine for food machines⁴⁰; Evanston $65 a machine⁴¹; in Ohio, Greene County requires a food license for each vending location unless it sells only prepackaged food that isn’t “potentially hazardous”, nuts, or wrapped gum or candy⁴².