Business Phone System Cost per User (VoIP)

Telephony is the category where the per seat model is both the most familiar and the least complete. Every other line of business software on this site bills you one number per user and stops. A phone system bills you a seat, then a fee attached to the seat, then a percentage of the seat, then whatever the seat did, then the object on the desk the seat talks through. Four of those five lines never appear on the pricing page, and together they are routinely worth more than the difference between two vendors you were agonising over.

This verdict prices a cloud phone system the way an invoice actually arrives. It walks the seat rungs and the specific capability that gates each jump, takes apart what unlimited means and where metered calling still exists, prices toll-free and international separately because they behave differently, works through the hardware fork and the buy against rent question, treats number porting as the schedule risk it is, explains the emergency service and regulatory recovery lines that never make the headline, separates contact centre seats from everything else, and closes with a modelled total at five, twenty-five and a hundred users. It sits alongside our verdict on the true cost of business software, which takes per seat billing apart in general, and our help desk software cost verdict, which prices the other tool your support queue usually runs on. Put your own seat count through the true-cost calculator and the companion on this page before you shortlist anything.

What a business phone system actually costs per user

Ask what a phone system costs per user and the honest answer has five parts, only one of which is advertised.

The first is the seat licence, the number on the pricing page, sold per user per month and scaling directly with headcount. The second is the per seat regulatory and emergency bundle, a small set of recurring charges attached to each seat that exist because telephony is a regulated service in a way that project management software is not. The third is usage, which covers everything falling outside the definition of unlimited in your plan, mainly toll-free inbound minutes and international outbound. The fourth is numbers, because direct dial numbers, toll-free ranges and additional lines are frequently priced individually rather than bundled per user. The fifth is hardware, which is one time, front loaded, and larger than most buyers expect.

Those five scale on different things. Licences scale with people. Fees scale with licences, which makes them invisible in a comparison and unavoidable in a bill. Usage scales with behaviour, which is the only line you can influence after signing. Numbers scale with how many external identities your business publishes. Hardware scales with how many people need something that is not a laptop.

The sixth part, the one that never appears anywhere, is the seat rung you were pushed into by a single requirement that one team had. Everything below takes those parts apart with illustrative planning figures, and the companion on this page reprices the whole stack for your headcount as you read.

The headline: illustrative per-seat pricing by rung

Start with the numbers people search for, framed as planning bands rather than quotes. Pricing in telephony moves constantly, varies by region and regulatory regime, and is heavily discounted against contract length and volume, so treat these as a shape to verify rather than a specification.

Entry seats commonly land around $18 per user per month. The job they do is give every person a number, a dial pad on a computer or a phone, voicemail, and a basic company greeting. Extension dialling, call transfer and a simple hunt group are usually present. What is not present is the machinery that turns a phone system into an operations tool.

Business seats commonly land around $26 per user per month. This is the most bought rung in the category. It typically adds a real auto-attendant with menus, call queues rather than simple ring groups, some form of call recording, integrations with common business software, and reporting that goes beyond a call log.

Advanced seats commonly land around $35 per user per month. The additions here are analytical and operational: richer historical reporting, supervisor visibility, deeper integration with a customer record, sometimes a first layer of queue management, and a contractual service commitment.

Enterprise seats commonly land around $45 per user per month and are bought for single sign-on, granular administrative roles, audit trails, device management at scale, a stronger service level agreement and the commercial terms that come with a negotiated contract.

Contact centre agent seats are not a rung on this ladder at all. They are a separate product, commonly starting around $95 per agent per month, and are covered in their own section below.

The useful thing about that chart is the bottom half. Nobody shops on a $1.50 emergency fee or a $5 number, and yet across a hundred seats and a handful of published numbers those two lines alone run into thousands a year. They are the reason effective cost per seat sits so far above sticker in this category.

What gates each jump between rungs

Rung names are marketing. The gates are consistent enough across the category to be worth naming, and knowing which gate you are about to cross is the difference between a considered upgrade and a surprise.

The first gate is the auto-attendant with real menus. Almost every entry plan gives you a greeting and a way to reach a person. Very few give you multi level menus, time of day routing, holiday schedules and different treatment for different published numbers. The moment somebody says the words press one for sales, you are usually shopping the second rung.

The second gate is call recording. Sometimes it is included at the business rung, sometimes it is an add-on at an illustrative $8 per user per month, and sometimes it is gated to advanced because the retention and search capability sits there rather than the recording itself.

The third gate is analytics that answer a question. A call log is not analytics. What people actually want is queue wait times, abandoned call rates, per person volumes over a period, and the ability to compare a week to the week before. That set commonly sits at the advanced rung.

The fourth gate is integration with the customer record, which is the single most common reason a business moves up a rung it did not plan to buy. The fifth is a contractual service level, which is bought at enterprise and which is worth reading very carefully.

Auto-attendant, menus and the first upgrade

The auto-attendant deserves its own section because it is where most small businesses meet the pricing model for the first time.

A greeting that says thank you for calling and then rings every phone in the building is a feature of nearly every plan. What a growing business actually wants is different: a menu that separates sales from support from accounts, a different path outside working hours, a holiday override that somebody non technical can set, and separate treatment for the number printed on the van versus the number on the website. Each of those is a routing decision rather than a greeting, and routing is what the second rung sells.

The cost question is not whether the feature is worth it. It usually is, because badly routed calls are expensive in a way that never shows on an invoice. The cost question is scope. Moving twenty-five people from an $18 seat to a $26 seat so that three published numbers can have menus is $2,400 a year of licence uplift, which is a real number to weigh against the alternative of a receptionist position or a simpler published number structure.

Ask every vendor whether menus and schedules are administered by a normal person in a web console or by the provider on a change request, because a routing change you have to raise a ticket for is a routing change you will stop making.

Call recording and the retention question behind it

Call recording is priced in three different shapes and the shape matters more than the rate.

The first shape is included at a rung, which sounds free and is not, because the rung applies to everyone. The second shape is a per user add-on, illustratively around $8 per user per month, which lets you attach recording to the people who need it and leave everyone else alone. The third shape is included capacity with paid retention, where recording itself is bundled but storage beyond a short window is billed separately.

That third shape is the one that surprises people. Recording generates data continuously, and any policy that says keep everything is a policy that grows a line item every month. Ask three questions before you compare recording prices: how long recordings are retained by default, what extended retention costs, and whether you can export recordings out of the platform in bulk if you leave.

Recording also carries obligations that have nothing to do with price. Notice and consent requirements for recording calls differ by country, by state or province, and sometimes by the type of call being recorded, and nothing on this page states what any of those rules require. If you are switching recording on, get the position that applies to your business and your callers from a qualified legal adviser before you configure anything.

Analytics, wallboards and what reporting is worth

Most buyers say they want reporting and then never open it, so it is worth being precise about the version that pays for itself.

Basic reporting tells you what happened: who called, when, how long. It is present nearly everywhere and it is close to worthless as a management tool, because it answers questions nobody was asking. Operational reporting tells you what is happening now: how many calls are waiting, how long the oldest one has waited, who is available. Analytical reporting tells you what changed: whether abandonment rose, whether Monday mornings need more cover, whether a particular published number is generating volume nobody planned for.

The second and third of those are what the advanced rung sells, and they only earn their premium if somebody owns the number. A wallboard on a wall that nobody looks at is a $9 per seat per month decoration. Across twenty-five people that is $2,700 a year.

The honest test is to name, before you buy, the one decision the reporting will change. Cover on a shift. A published number retired. A queue split in two. If nobody can name a decision, buy the cheaper rung and revisit in six months. Our verdict on running a software trial covers how to test that claim rather than accept it.

CRM integration and the click-to-call trap

Integration with the customer record is the most common unplanned upgrade in telephony, and it arrives in four levels of depth that are easy to confuse.

The shallowest is click-to-call, where a number in another application dials through your phone. It is genuinely useful, it is often available at a low rung, and it is not integration in any meaningful sense. Above that is screen pop, where an inbound call surfaces the matching customer record. Above that is activity logging, where calls, durations and outcomes are written back automatically against the record. At the top is embedded telephony, where the whole dial pad and call controls live inside the other application and the phone system becomes invisible.

Vendors describe all four as integration. The price difference between the shallowest and the deepest is typically a full rung, and sometimes a connector charge on top. Before you pay for depth, decide which level solves your actual problem, because activity logging is what most teams want and embedded telephony is what most teams are sold.

Our verdict on what a CRM really costs covers the same question from the other side, since the customer platform often charges for the telephony connector too. Two vendors both billing for the same integration is a normal and avoidable outcome.

SLA, uptime credits and what enterprise really sells

The enterprise rung is bought for three things, and only one of them is technical.

The first is identity and administration: single sign-on, provisioning from a directory, granular admin roles, audit logging. The second is device and estate management at a scale where somebody has to remotely configure hundreds of handsets. The third is contractual, and it is the one that gets waved around in a sales meeting: a service level agreement with a stated availability figure and a credit if it is missed.

Read the credit clause rather than the number. Availability commitments in cloud services are typically measured monthly, typically exclude scheduled maintenance, and typically pay out as a percentage credit against the following month rather than as compensation for your losses. A credit worth a fraction of one month of licences is not insurance against a day without phones. It is a signal about how seriously the provider takes its own uptime, and it should be read as a signal.

If phones going down would genuinely stop your business, the money is better spent on the failover design than on the rung: numbers that overflow to mobiles automatically, a second internet path, and a written plan somebody has actually tested. That is an architecture question rather than a pricing one, and no rung buys it for you.

Metered versus unlimited calling

Unlimited is the word that does the most work on a telephony pricing page, and it is a scope statement rather than a promise.

In practice unlimited normally means unmetered outbound calling to ordinary domestic numbers, from a device in the country the plan belongs to, at a volume consistent with office use. It normally excludes premium rate ranges, certain special service numbers, directory services, most international destinations, and inbound minutes on toll-free numbers. Many plans carry an acceptable use clause that reserves the right to review accounts whose calling pattern resembles an auto dialler or an outbound campaign.

Metered plans still exist and are still sometimes the right answer. A business where most people receive rather than place calls, or where a large share of seats are occasional users, can pay a lower seat rate plus an illustrative $0.018 a minute domestic and come out ahead. The arithmetic is simple: the break-even between an unlimited seat and a metered seat priced a few dollars lower is a fixed number of minutes a month, and most offices are nowhere near it.

The trap is buying unlimited for reassurance and then discovering that the minutes you actually worried about, the international ones and the toll-free ones, were never in scope.

Toll-free numbers and the inbound minute

Toll-free numbers invert the usual model, and that inversion is the whole reason they need separate treatment.

On an ordinary number the caller pays for the call and you pay nothing to receive it. On a toll-free number you pay for the call the caller placed. So a toll-free number has two prices: a recurring charge for holding the number, illustratively around $15 a month, and a per minute inbound charge, illustratively around $0.022 a minute, that scales with how successful your marketing is.

That second characteristic is the one to model honestly, because it means a good month costs more than a bad one. A number receiving two thousand inbound minutes a month at that illustrative rate is $44 in usage on top of the $15 rental, which is trivial. The same number receiving forty thousand minutes because a campaign worked is $880 a month, which is not.

Three practical checks. First, ask whether inbound toll-free minutes are ever included in an unlimited plan, because occasionally a bundle exists. Second, ask what happens to the rate at volume, since toll-free pricing is frequently tiered downward. Third, ask whether the platform can cap or alert on toll-free spend, because an uncapped inbound number is the one line on a phone bill that can move by an order of magnitude without anyone changing a setting.

International calling and the rates nobody models

International rates are the least modelled and most variable line in a phone bill, and any single number quoted for them is misleading.

The structure is what matters. International outbound is priced per minute per destination, and the spread between destinations is enormous: a landline in a neighbouring country and a mobile in a distant one can differ by two orders of magnitude. Calls to mobile numbers are routinely more expensive than calls to landlines in the same country, sometimes several times more, because a mobile termination charge sits behind them. Some destinations and some number ranges within otherwise ordinary destinations carry premium rates that exist specifically to catch traffic.

Because of that spread, the only useful exercise is a specific one. Take your last three months of call records, group the international minutes by destination, and ask each vendor to price that specific list. A blended average is meaningless when eighty percent of your international spend goes to one country.

Two structural options are worth asking about. Some providers sell an international add-on bundle, a fixed monthly amount per seat covering a named set of destinations, which is worth it only for people who genuinely call abroad every day. Others sell a local number in the destination country, which converts an international call for your customer into a domestic one for them, and is frequently the cheaper answer to the underlying problem.

Numbers, direct dial and the per-number line

Numbers are the quiet multiplier in telephony pricing and the line most often missed in a comparison.

The base assumption in most plans is one number per seat, so a twenty-five seat account gets twenty-five direct dial numbers and a main company number. Everything beyond that assumption is usually priced individually at an illustrative $5 per number per month. Businesses accumulate numbers faster than they realise: a number per department, a number per location, a number per marketing campaign so that response can be attributed, a number kept alive because it is printed on something, a fax number nobody will admit to needing.

Thirty extra numbers at that illustrative rate is $1,800 a year, which is roughly the same as adding six business seats. Nobody makes that decision consciously, because each number was added on its own and approved by nobody.

Two questions worth asking. First, how many numbers are included per seat and whether unused seat numbers can be pooled. Second, what happens to numbers you stop using: whether they are billed until explicitly released, which is the usual answer, and whether releasing them is self service. An audit of published numbers is one of the highest yield hours available in this category and it costs nothing.

The hardware fork: desk phones, softphones and headsets

Hardware is the largest single one time line in most deployments and the one where the decision is genuinely yours rather than the vendor’s.

Softphone only costs nothing in hardware. The application runs on a computer or a mobile device and the call quality depends on whatever audio path the person already has. It works well for people who take a handful of calls a day and badly for anyone who lives on the phone.

Headset plus softphone, illustratively around $60 for a serviceable wired unit and around $180 for a wireless one that someone can wear all day, is the middle path and the one most offices land on. The illustrative $90 a head used in the model here reflects a mix of the two.

Desk phones, illustratively around $120 for an entry model, around $220 for a mid range unit with a colour display and multiple line keys, and around $380 for a video or executive model, are worth buying where a physical handset is genuinely required. Reception positions, warehouses, workshops, shared spaces, conference rooms and any role where a call must be answered without a laptop being awake all justify one.

The mistake is buying a desk phone for every seat by default. On twenty-five people that is $5,500 against $2,250 for headsets, a $3,250 difference that arrives entirely in your first year. Decide by role rather than by policy.

Buy the handsets or rent them

Once you have decided somebody needs a physical phone, the second decision is whether it sits on your balance sheet or the provider’s.

Buying means the illustrative $220 lands once. You own the device, you can take it to another provider if it is a standard handset and the new provider supports it, and you carry the risk of it failing outside warranty. Across a three year term, buying is almost always the cheaper cash total.

Renting, illustratively around $6 per phone per month, converts that $220 into $216 across three years, which looks like a wash until you account for the two things rental actually buys: replacement of failed units without an argument, and no capital outlay in a first year that is already front loaded. It also means the phone is not yours, which matters at the end of a term.

The number to compute is the crossover. At an illustrative $6 a month, a $220 phone pays for itself in about thirty-seven months, so rental is worse on cash beyond about three years and better inside it. If your contract is three years and you expect to re-evaluate at the end, rental is defensible. If you expect to keep the estate for five, buying is clearly cheaper.

The version to avoid is rental with a locked handset that only works on that provider’s platform, because it quietly converts a hardware decision into a switching cost. Our verdict on negotiating software pricing covers how that kind of lock-in shows up at renewal.

Number porting and how long it really takes

Porting is the schedule risk in a phone migration, and treating it as an administrative formality is how cutover dates get missed.

The mechanics are straightforward to describe and fiddly to execute. You authorise the new provider to request your numbers, they submit the request to the carrier currently holding them, that carrier validates the request against its own records, and a cutover window is scheduled. The failure modes are almost always clerical: a company name that does not match the one on the old account, a service address that was updated in one system and not the other, an account number transposed, an outstanding balance, or a number that turns out to be tied to a physical circuit rather than to your account.

Timescales vary by country, by carrier and by the size and type of the range being moved, and the rules governing them are set by regulators rather than by vendors. A small port of a handful of ordinary numbers is typically measured in weeks rather than days. A large range, a number tied to a line, or an account with any discrepancy takes longer. Confirm the expected window in writing with both carriers rather than working from any general figure, including this one.

The practical protection is sequencing: run the new system in parallel on temporary numbers, prove it works with real calls, and port only when the platform is otherwise finished. Our verdict on migrating to new software covers the parallel running principle in general.

E911 and the fees that follow a street address

Emergency calling is the structural reason a phone bill looks different from a software bill.

A fixed line has a location by definition. An internet phone does not, because the device can be moved and the account can be used from anywhere. To make emergency calling work, the platform has to hold a registered service address for each seat, route emergency calls to the correct answering point for that address, and pass location information along with the call. Some platforms also notify an internal contact when an emergency call is placed from a company device.

Providers recover the cost of that capability as a small recurring per seat charge, illustratively around $1.50 a seat a month in the model used here. Across a hundred and eight seats that is $1,944 a year, which is not a rounding error even though it looks like one on a single line.

The obligations behind the fee are the part that matters more than the money. Requirements for registering addresses, keeping them updated when people move desks or work from home, notifying internal contacts, and displaying information to users differ by country and frequently by state or province, and nothing here states what any of those rules require for you. Ask each provider how addresses are captured and updated, whether updating one is self service, and what happens when a registered address is missing. Then confirm your own obligations with the provider and a qualified adviser.

Regulatory recovery fees and the gap from the headline

The line that generates the most confusion on a first invoice is the one computed as a percentage rather than a rate.

Telephony providers commonly levy a charge described as regulatory recovery, cost recovery, compliance and administrative, or some combination of those words. Structurally it is a surcharge applied to the licence subtotal rather than a per seat amount, and in the model used here it is an illustrative nine percent. Some providers itemise it, some fold it into a line called fees and surcharges, and some quote a plan price that includes it, which makes cross vendor comparison harder than it should be.

Two honest points about it. First, part of what sits behind such charges genuinely is imposed on the provider by regulation, and the mix differs by jurisdiction; the exact composition and whether any portion is a pass through or a margin is something only the provider can answer for its own invoice. Second, because it is a percentage, it scales with everything you buy. Moving from a $26 seat to a $35 seat does not just cost $9, it costs $9 plus the surcharge on $9.

The practical move is to ask every vendor for an all in per seat figure at your headcount, in writing, including all recurring surcharges. A vendor who produces that number readily is one you can budget for. A vendor who insists on quoting the plan price alone is telling you something about the invoice you will receive.

Contact centre seats are a different product

The single most expensive misunderstanding in telephony pricing is treating a contact centre seat as a premium phone seat.

A unified communications seat gives a person an identity on the phone network: a number, a dial pad, voicemail, presence, usually meetings and messaging. A contact centre seat gives an organisation control over a queue: routing rules based on skill, priority and time, agent states, supervisor listen and whisper, real time queue metrics, historical analysis at interval level, quality management and scoring, and often workforce scheduling and forecasting.

That is a different product with different buyers, and it is priced against the value of running a queue rather than the cost of a line. An illustrative $95 per agent per month is roughly three and a half times a business seat.

The consequence for budgeting is simple and frequently ignored. Count the people who work a queue all day, and buy contact centre seats for exactly those people. Everybody else gets an ordinary seat. A hundred person company with eight queue workers should be buying eight agent seats, not a hundred. The difference between those two on illustrative figures is over $88,000 a year, which is larger than the entire rest of the phone bill.

If your support team already runs on a ticketing platform, check what its own voice module costs before adding a contact centre from the phone vendor. Our help desk software cost verdict prices that alternative.

Seat minimums, terms and the annual prepay question

Commercial structure in telephony is closer to a telecoms contract than to a software subscription, and it behaves accordingly.

Seat minimums are common at the upper rungs, which means a small team buying an enterprise seat for one single sign-on requirement can find itself paying for more seats than it has people. Contract terms of one to three years are normal, and the discount for length is real and often substantial. Annual prepay discounts sit on top of that in many cases.

The trade is straightforward and worth naming. A longer term buys a lower rate and sells you an exit. Telephony is stickier than most software because of numbers, hardware and the fact that a failed migration is audible to every customer, so the switching cost is already high before a contract adds to it. Signing three years for a meaningful discount is defensible. Signing three years for a small one is paying for the privilege of not being able to leave.

Two clauses to read specifically. What happens to your rate at renewal if you do nothing, since uplift language is common and frequently uncapped. And whether seat counts can go down mid term or only up, because a ratchet clause turns a headcount reduction into a bill that does not move. Our verdict on cancelling a software subscription covers how those clauses tend to be written.

Implementation, provisioning and the network work nobody quotes

Setting up a cloud phone system is cheap in professional services and expensive in attention, and the split is different from other software.

The vendor side is usually modest. In the model here, one time setup is an illustrative $300 base plus $20 a seat, covering account configuration, number provisioning, routing build and handset registration. On twenty-five seats that is $800. Some vendors bundle it, some sell it as a package, some hand you a self service console and a video.

The side nobody quotes is your own network. Voice is the least forgiving traffic on an office connection because it is real time: a video call degrades visibly and recovers, a voice call breaks in a way people remember. Whether your existing connection, switches and wireless can carry it is a question that gets answered during the trial or during the first bad week. Where a site needs cabling to desks that never had a phone, or a switch that can power handsets, that is a real cost sitting outside every quote you will receive.

The other unquoted line is the routing design itself. Deciding what happens to a call at every hour of every day, who covers what, and where the overflow goes is a business conversation rather than a technical one, and teams that skip it configure the same thing three times.

What the bill looks like at five, twenty-five and a hundred users

Three illustrative businesses, three configurations, three very different totals. Every figure below is a planning number chosen to show how the model behaves, and the same arithmetic runs live in the companion on this page.

Five people, business seat, headsets, no queue. Licences at $26 are $130 a month. Emergency service at $1.50 a seat is $7.50. Regulatory recovery at nine percent of licences is $11.70. An illustrative hundred chargeable minutes of toll-free and international at a blended $0.05 is $5. Monthly total: $154.20, so $1,850.40 a year. Headsets at $90 a head are $450 and setup at $300 plus $20 a seat is $400, so $850 one time. First year: $2,700.40, which is about $45.01 per seat per month against a $26 sticker.

Twenty-five people, business seat, desk phones, no queue. Licences are $650 a month, emergency service $37.50, recovery $58.50, and six hundred chargeable minutes $30. Monthly total: $776, so $9,312 a year. Desk phones at $220 are $5,500 and setup is $800, so $6,300 one time. First year: $15,612, about $52.04 per seat per month.

A hundred seats plus an eight agent support desk, advanced seat, headsets. Licences are $3,500 for the hundred plus $760 for the eight agent seats, so $4,260. Emergency service across a hundred and eight seats is $162, recovery is $383.40, and twenty-five hundred chargeable minutes are $125. Monthly total: $4,930.40, so $59,164.80 a year. Headsets across a hundred and eight seats are $9,720 and setup is $2,460, so $12,180 one time. First year: $71,344.80, about $55.05 per seat per month.

Why the per-seat curve stops flattening

In most software categories the effective cost per user falls as you grow, because fixed costs spread. Telephony behaves differently, and the three worked examples show it: $45.01, then $52.04, then $55.05 per seat per month across a first year, rising rather than falling.

Three structural reasons sit behind that. Hardware scales one to one with people rather than spreading, so a hardware heavy deployment never dilutes. Contact centre seats appear at scale and not before, and they are nearly four times an ordinary seat. And rung requirements arrive with size, so a hundred person company is usually on an advanced seat where a five person company is on a business one.

The counterweight is negotiation, which genuinely does move at volume. Discounts against list, longer terms, bundled hardware and waived setup all become available in a way they are not for a five seat account, and a hundred seat deal that pays list has not been negotiated. None of that is modelled in the figures above, which are deliberately list based so you can apply your own discount to them.

The lesson is not that scale is bad. It is that scale is where the negotiation returns exceed the effort, and that the per seat number you were quoted at five people is not a floor you will drift down towards. Run your own counts through the true-cost calculator and the companion to see where your version of that curve sits.

Signs your phone bill has drifted

Phone bills drift more quietly than software bills because the invoice is longer and less legible. Five signals are worth checking annually.

Seats outnumber people. Telephony seats are rarely reclaimed when somebody leaves, because the number stays live and nobody wants to break a routing rule they do not understand. Reconcile the seat list to the payroll list once a year.

Numbers nobody can account for. Pull the full number inventory and ask, for each one, what publishes it and what happens when it rings. Numbers with no answer to both questions are pure cost.

Contact centre seats on non queue workers. Agent seats get assigned during a rollout and never reviewed. At nearly four times an ordinary seat, a handful of stragglers is a meaningful annual number.

Toll-free usage moving without a campaign. A rising inbound minute line with no marketing behind it usually means a number is being used for something it was not intended for.

A renewal rate that moved on its own. Uplift clauses are common and quiet. Compare your current effective per seat rate to what you signed, not to the pricing page.

What to ask a phone vendor before you sign

Nine questions, all answerable in writing, all of which change the number.

What is the all in per seat cost at my headcount, including every recurring surcharge? This is the only figure worth comparing across vendors.

What exactly is excluded from unlimited? Ask for the list, not the adjective.

Price my actual international destinations. Supply three months of grouped call records rather than asking for a rate card.

How many numbers are included, and what does an additional one cost?

Is call recording included, add-on or retention limited, and what does extended retention cost?

Which specific capabilities are gated to the rung above the one you are quoting me?

Are handsets locked to your platform, and what is the rental crossover against purchase?

What is the porting process, what could delay it, and what window are you committing to?

What happens to my rate at renewal if I do nothing, and can seat counts go down mid term?

The bottom line

A business phone system is priced like software and billed like telecoms, and almost every unpleasant surprise in this category comes from that mismatch. The seat rate is real, it is a minority of the bill, and it is the only part most buyers compare.

The parts that decide your actual number are structural rather than promotional. Whether people need handsets, and whether you buy or rent them. Whether the minutes you care about are inside or outside the definition of unlimited. How many numbers you publish and how many you have forgotten. How many people genuinely work a queue. And the percentage surcharge that sits on top of everything else and moves whenever anything else moves.

On the illustrative figures modelled here, a $26 seat becomes about $45 a seat a month for five people, $52 for twenty-five with desk phones, and $55 for a hundred running an eight agent queue. None of those are quotations and all of them are the same lesson: the sticker is roughly half the story, hardware is the largest thing you control, and the contact centre seat count is the single decision with the most money attached.

So do the boring parts first. Count the queue workers separately. Audit the numbers. Decide handsets by role rather than by policy. Get an all in per seat figure in writing from every vendor. Then negotiate, because in this category the list price is a starting position rather than a price, and the buyers who know what is on the invoice are the ones who get it moved.

VetLoft is funded by its readers rather than by the vendors it prices, and this verdict is written on that footing: educational material only, not procurement, telecoms, regulatory or legal advice for your organisation. Every seat rate, surcharge percentage, per minute figure, handset price, setup estimate and worked total above is an illustrative planning number chosen to show how per seat telephony billing behaves, not a quotation from any provider, and pricing in this category is discounted, regionalised and revised often enough that a figure typical when this was written may read differently by the time you shop. Which capability sits at which rung varies by provider and shifts between releases, so read the ladder here as a shape to verify rather than a specification. Emergency calling obligations, service address registration, call recording notice and consent, number portability timescales and the composition of regulatory surcharges are set by rules that differ by country and often by state or province, and nothing above states what any of those rules require of you; obtain the position applying to your business from the provider and a suitably qualified adviser before configuring or switching on anything. Confirm seat definitions, minimums, inclusion scope, number allowances, retention, porting windows, handset ownership, export rights and renewal uplift directly with each provider in writing before relying on any figure here.