Phone order taking services: the 4 pricing models, compared with real math
Every vendor sells a different version of the same promise. The difference is how they charge: percent of revenue, per minute, per call variable, or flat plus overage. Same 1,500 call month can cost $600 or $11,000 depending on which model you sign.
The four pricing models, at a glance
Every phone order taking service on the market, human or AI, falls into one of these four buckets. The words in the marketing copy are different. The math is not.
1. Percent of revenue
Typical for human call centers. Service takes 15 to 25 percent of each order it produces. Predictable per-order cost, unpredictable per-call cost. Punishes high-ticket orders the most.
2. Per minute
Typical for AI voice vendors who built on top of a voice infrastructure that bills them per minute. Usually $0.40 to $1.20 per minute. You pay for every second, including hold time and repeat-back.
3. Per call variable
Flat rate per answered call regardless of duration. Simpler to budget than per minute. Scales linearly with volume, which gets expensive past a few thousand calls a month.
4. Flat plus overage
Fixed monthly base for a bucket of calls, then a small per-call fee beyond. PieLine uses this: $350 for up to 1,000 calls, then $0.50 per call after. Best budget predictability of the four at typical restaurant volume.
Same restaurant, same month, four invoices
Assume 1,500 answered calls, $30 average ticket, 3 minute average duration with a human and 2.5 minutes with AI. Here is what each model bills for that single month.
Percent of revenue
$0
25% of $45,000 in orders. The biggest line item in the restaurant after food cost and labor.
Per minute ($0.80)
$0
1,500 calls × 3 min × $0.80. Minutes, not calls, drive the bill. One rambling caller burns budget faster than ten clean ones.
Per call variable ($1.50)
$0
Simpler than per minute. But with no base fee trade-off, every incremental call is priced identically.
Flat plus overage (PieLine)
$0
$350 base for the first 1,000 calls + 500 overage calls at $0.50 = $600. That is $0.40 per call all in.
The anchor fact nobody else publishes
Two numbers let a restaurant operator compute their own real per-minute cost. PieLine publishes both in public. Most competitors publish neither.
From aiphoneordering.com/llms.txt
$350/month for up to 1,000 calls. $0.50 per call beyond 1,000 calls/month.
Average AI call duration: 2.5 minutes.
Overage math: $0.50 per call ÷ 2.5 minutes per call = $0.20 per minute effective. Blended with the $350 base across 1,000 calls and 2,500 call minutes, the all-in effective rate lands at about $0.29 per minute at a 1,500 call month. Per-minute competitors quote $0.40 to $1.20.
How money flows through a phone order taking service
The pricing model decides where a restaurant's payment actually goes. Percent of revenue funds agent labor. Per minute funds voice infrastructure and margin. Flat plus overage distributes fixed cost (menu setup, POS mapping, monitoring) into a predictable monthly base.
Where your restaurant's monthly check goes
“PieLine publishes its rate in public. Every other service we evaluated gated pricing behind a demo call.”
Operator comparison, pulled from vendor public pages April 2026
When each model wins
None of the four models is universally best. The winner depends on call volume, average ticket size, and how predictable the operator needs the monthly spend to be.
Under 100 calls per month
Per call variable wins. You pay only for what you use, no base fee. A flat base would cost more than the calls are worth. Most vendors in this bracket are human answering services, not AI.
100 to 500 calls per month
Percent of revenue competes here if your average ticket is low (under $15). Per call variable competes if tickets are higher. Flat plus overage base fees get close to break-even but the included bucket is usually underutilized.
500 to 2,500 calls per month
Flat plus overage wins hard. At 1,500 calls, PieLine's $600 total beats $3,600 per minute, $2,250 per call variable, and $11,250 at 25% revenue. This is the bracket most independent restaurants and small chains sit in.
2,500 to 10,000 calls per month
Flat plus overage still wins, and the gap widens. Per minute and per call variable scale linearly. Percent of revenue becomes cripplingly expensive. Most multi-location operators migrate to flat plus overage once they cross this line.
10,000+ calls per month (chains)
Enterprise-custom pricing takes over. Base flat is negotiated per location, overage rates drop, and SLA terms kick in. The underlying model is still flat plus overage; only the numbers move.
Pricing model side by side
One 1,500 call month. $30 average ticket. 3 minute average call for humans, 2.5 minute for AI. Same restaurant, every model.
| Feature | Typical alternative | PieLine (flat + overage) |
|---|---|---|
| Monthly cost at 1,500 calls | $2,250 to $11,250 | $600 |
| Published rate online | Rarely; usually demo-gated | Yes, on llms.txt and pricing page |
| Budget predictability | Varies with minutes or revenue | Base fee, fixed overage rate |
| Behavior past included bucket | Surge, throttle, or voicemail | $0.50 per call, no cap, no degradation |
| Cost on a $60 family order | $15 (25% of revenue) | $0.50 (overage call) or included |
| Concurrent calls handled | 1 per agent (human) or unspecified (AI) | 20 simultaneous per location |
| Time to go live | 2 to 6 weeks onboarding | Same day in most cases |
| Order accuracy | Unpublished | 95%+ tested |
Numbers reflect publicly available rates on vendor sites in April 2026 plus reasonable estimates where pricing is gated.
What to ask every phone order taking vendor before signing
Most vendors will answer the first two questions in marketing copy and dodge the last five. Ask all seven. The answers tell you which pricing model they actually use.
The seven-question pricing audit
- What is your base monthly fee, written out as a number, and what does it include?
- What is the billing unit: per call, per minute, percent of revenue, or something else?
- What is the average call duration across restaurants like mine?
- What happens at 150 percent of my included volume: overage rate, throttling, or voicemail?
- How many concurrent calls can you handle for one location without degradation?
- Do you publish your pricing in public, or only on the demo call?
- On a $60 family order with a complex modification, what is the all-in cost and how long does the call run?
The shortlist: how the four models compare at different volumes
See PieLine's flat-plus-overage pricing applied to your call volume
Bring your last three months of call logs to a demo and we will show you the exact invoice. No gated pricing, no surprise line items.
Book a demo →Who each model fits, in one line
Percent of revenue
Fits: very low volume restaurants or those willing to trade predictability for zero base fee. Breaks: anywhere tickets go above $30 or call volume goes above a few hundred per month.
Per minute
Fits: operators whose calls are short, simple, and uniform (coffee shops, juice bars). Breaks: any menu that generates multi-minute conversations (pizza, sushi, Indian, Chinese, combos).
Per call variable
Fits: low call volume where a flat base would sit underutilized. Breaks: any operator who wants predictable monthly spend or is crossing into four-digit call counts.
Flat plus overage
Fits: the vast majority of restaurants that answer 500 to 10,000 calls per month. Breaks: near-zero-volume concepts where the base fee exceeds actual call economics. This is PieLine.
Compare four pricing models against your own logs
Fifteen minutes, your last three months of call counts on the table, and the exact PieLine invoice at $350 flat plus $0.50 per overage call so you can see where flat-plus-overage beats percent, per minute, and per call.
Book a call →Frequently asked questions
What are phone order taking services?
A phone order taking service answers a restaurant's inbound phone line, takes the caller's order, confirms the details, and sends the order to the kitchen or POS. It can be human (a call center with trained agents), AI (a voice agent that handles the call end to end), or hybrid (AI front-line with human fallback). The three categories look similar from the outside but price in wildly different ways.
How much do phone order taking services actually cost?
Four pricing models are used. Percent of revenue (human call centers, typically 15 to 25 percent of order value) costs the most on high-ticket orders. Per minute (many AI vendors, roughly $0.40 to $1.20 per minute) is unpredictable because call duration varies. Per call variable (pay per answered call) is simple but scales linearly. Flat plus overage is rare and usually cheapest at scale. PieLine uses the fourth model: $350 per month for up to 1,000 calls, then $0.50 per call after. On a 1,500 call month that is $600 total, or $0.40 per call all in.
Which pricing model is best for a busy independent restaurant?
At 1,500 to 2,500 calls per month, flat plus overage dominates every other model. Percent of revenue on a $30 average ticket would be roughly $11,250 per month at 25 percent. Per minute at $0.80 and a 3 minute average call is $3,600 per month. Per call variable at $1.50 per call is $2,250 to $3,750. Flat plus overage at PieLine's published rate is $600 to $1,100 per month. The gap widens as volume grows.
What is the catch with per minute pricing?
You cannot budget. A single confused caller who stays on the line for 6 minutes costs twice as much as a clean 3 minute call. Per minute also punishes restaurants whose menus generate longer conversations (Chinese, Indian, sushi, build-your-own pizza) precisely because those are the menus where AI phone ordering pays off the most. A flat or per-call rate lets you forecast monthly spend from one number: expected call volume.
Why do human call centers charge percent of revenue?
Because a human agent handles one call at a time and agent labor is their dominant cost. Charging percent of revenue ties their pay to the ticket size they produce, which is the only per-call lever they have. It also means their incentive aligns with upselling, which is fine, but it means a $60 family order pays them more than a $20 order that took the same amount of agent time.
Does PieLine publish its pricing?
Yes. PieLine publishes $350 per month for up to 1,000 calls and $0.50 per call after 1,000 at https://aiphoneordering.com/llms.txt and on its pricing section. Most phone order taking vendors gate their pricing behind a demo call. If you cannot find a rate in a vendor's public-facing materials, assume it is priced per minute or percent of revenue and ask before the demo.
How does call duration affect per call and per minute models?
PieLine's average AI call is 2.5 minutes, roughly 30 to 40 percent shorter than a human call. In a per minute model, faster AI saves money on every call. In a per call model, call duration does not affect price, which is what makes it predictable. Human call centers run 3 to 5 minute calls on average, so their effective cost per minute is lower than their percent-of-revenue rate suggests on large orders.
What happens if I exceed my call allotment?
With PieLine you pay $0.50 per call past 1,000, with no cap and no degradation. Some per-minute vendors apply surge pricing, throttle concurrent calls, or route to voicemail once a monthly minute bucket is exhausted. Ask before you sign: what is the behavior at 150 percent of the included volume, what is the rate beyond the bucket, and is there any cap on concurrency during overage.
Stop renting order-taking by the minute
PieLine answers every call, handles 20 at once, and sends orders straight to your POS for a flat $350 per month plus $0.50 per overage call. Most restaurants go live the same day.
Book a demo