The Cost of Missed Calls: How to Measure Revenue Lost to Unanswered Inbound Calls

A cost that never appears in a report

Every other channel accounts for itself. Abandoned baskets are counted, bounced sessions are measured, unreplied emails sit visibly in a queue. An unanswered phone call leaves nothing behind except a line in a telephony log that nobody opens, and the caller who rang a competitor instead is recorded nowhere at all.

This produces a predictable organisational blind spot. Teams debate conversion rate optimisation at length while a measurable share of their highest-intent traffic — people who picked up a phone — reaches a ringing tone. The reason is not indifference; it is that nobody has put a number on it, and unnumbered problems lose to numbered ones.

This article is a method for producing that number from your own data. It deliberately contains no industry benchmarks. A borrowed percentage is easy to dismiss and usually wrong for your business; a figure derived from your own telephony logs is neither.

What counts as a missed call

Before measuring, define the categories. Most telephony systems report these differently and conflating them produces a number nobody trusts.

1. 1

   Unanswered during opening hours

   The phone rang, nobody picked up, the caller hung up. The clearest category and usually the one that causes the most internal discomfort, because it happened while the business was open.

2. 2

   Abandoned in queue

   The caller reached a queue or a hold and left before being answered. Worth separating because the fix is different — this is a capacity and wait-time problem rather than an availability one.

3. 3

   Busy or blocked

   All lines occupied, so the call never entered a queue. Frequently invisible in reporting because the call did not reach the system that does the reporting.

4. 4

   Out of hours

   Rang outside opening times, reached voicemail or nothing. Usually the largest single category for businesses selling to consumers or across time zones, and the one most easily dismissed as unavoidable.

5. 5

   Voicemail left but never returned

   A distinct and embarrassing category. The caller made the effort; the business did not close the loop.

6. 6

   Answered but abandoned as a lead

   Someone picked up, took a message, and nothing happened. Not a missed call technically, and identical to one commercially.

The calculation framework

The method has five inputs. Four come from systems you already have; one is a judgement that should be made conservatively and documented.

1. 1

   Input 1: missed call volume, by category and by hour

   From your telephony provider. Break it down by hour of day and day of week — the pattern matters more than the total, because it tells you what kind of intervention would help.

2. 2

   Input 2: the share that were genuine prospects or customers

   Not every missed call is a lost opportunity. Some are suppliers, recruiters and wrong numbers. Estimate this from a sample of answered calls over the same period, classified by who was calling and why. Using answered calls as the sample is the defensible approach, because it is real data about your own call mix.

3. 3

   Input 3: your own conversion rate for answered calls

   From your CRM: of inbound calls from prospects, what share became a customer or a qualified opportunity? This is the number that makes the calculation yours rather than borrowed.

4. 4

   Input 4: your average order or contract value

   Also from your own records. Use the median rather than the mean if a few large deals distort it, and say which you used.

5. 5

   Input 5: the recovery assumption

   The share of missed callers who would have converted had the call been answered. This is the only genuinely uncertain input, and it must be conservative. Many callers ring several suppliers; some would have been lost regardless. Pick a deliberately cautious figure, state it explicitly, and show the result at two or three different assumptions rather than one.

Multiply: missed calls from genuine prospects, times your own answered-call conversion rate, times your own average value, times the recovery assumption. Present it as a range across assumptions, never as a single confident number. A range that a sceptical finance director can interrogate is far more persuasive than a precise figure they can dismiss.

Reading the pattern, not just the total

The total justifies attention. The distribution tells you what to actually do, and the two conclusions are often very different.

• Missed calls concentrated in known peaks — lunchtime, Monday mornings, after a campaign — indicate a capacity problem. The answer may be scheduling or overflow handling rather than technology.
• Missed calls spread evenly through opening hours suggest chronic under-staffing relative to volume, which is a resourcing decision.
• A large out-of-hours share points at coverage, and it is usually the category with the clearest business case because the alternative is nothing at all.
• High queue abandonment with low unanswered volume means calls are being answered but not quickly enough — a wait-time problem with different remedies.
• Busy or blocked calls suggest a line capacity issue that is often trivially cheap to fix and frequently goes unnoticed for years.
• Voicemails not returned is a process failure rather than a capacity one, and it is the cheapest item on this list to correct.

Two of these — line capacity and unreturned voicemails — usually cost very little to address and should be fixed before any larger programme is considered. Presenting them alongside the bigger number makes the analysis credible rather than self-serving.

Second-order costs worth naming

The direct revenue calculation understates the position, and these effects are worth stating qualitatively even where you cannot put a figure on them.

• Marketing spend that generated the call is wasted in full, which means the effective cost per acquisition on phone-driven campaigns is higher than reported.
• Existing customers who cannot get through are a retention risk, not a sales one, and that cost lands in a different report entirely.
• Repeat attempts consume capacity: a caller who rings three times generates three call records and one opportunity, which distorts volume reporting.
• A competitor answering first frequently sets the terms of the comparison, which matters most in sectors where buyers contact several suppliers.
• Staff interrupted mid-task to answer overflow do both things worse, a cost that never appears anywhere.
• Reputation effects where the business is visibly hard to reach.

Name these without quantifying them unless you genuinely can. An analysis that is precise where it can be and honest about what it cannot measure carries more weight than one that puts a number on everything.

What to do with the number

The measurement is only useful if it leads to a decision. Map each pattern to the intervention that addresses it.

1. 1

   Line capacity and unreturned voicemails

   Fix first. Low cost, immediate effect, and doing so demonstrates the analysis was not a pretext for a larger purchase.

2. 2

   Peak-hour overflow

   Options are scheduling changes, an overflow destination, or automated handling of the calls that do not need a person. Which is right depends on what those peak calls actually are — sample them before deciding.

3. 3

   Out-of-hours coverage

   The realistic comparison is not automation against a person; it is automation against a ringing tone. Capture, qualification and a booked next-day commitment is usually achievable, along the same lines as any after-hours flow, applied to the voice channel.

4. 4

   Queue abandonment

   Either reduce wait or change what happens during it — a callback offer that holds the caller's place converts an abandonment into a scheduled conversation.

5. 5

   Routine calls consuming capacity

   Where a large share of answered calls are status checks, opening hours and other documented questions, handling those automatically returns capacity to the calls that need a person. This is the argument for an AI call centre that is actually supported by the data rather than assumed.

The honest version of this analysis frequently concludes that part of the problem is process and part is capacity, and that only one slice justifies new technology. That conclusion is more credible and tends to get funded.

Data requirements

The analysis needs less than teams expect, but the pieces have to be genuine.

• Telephony reporting with the six categories separable, hour by hour, over at least a quarter.
• A sample of answered calls classified by caller type and purpose — a week of manual classification is usually enough.
• CRM data linking inbound calls to outcomes, which is the input most often missing and the one that makes the calculation defensible.
• Average or median deal value from your own records.
• Marketing spend attributable to phone-driven campaigns, if you want the wasted-spend figure.
• A documented, conservative recovery assumption with the reasoning behind it.

If calls are not linked to outcomes in your CRM, that is worth fixing before anything else — without it, every conversation about phone channel performance remains an opinion. Our automation services page covers how call data is typically connected to CRM records.

Common mistakes in the analysis

This calculation is easy to do in a way that nobody believes.

• Using an industry benchmark instead of your own conversion rate, which invites the response that your business is different.
• Assuming every missed call was a prospect.
• Assuming every missed prospect would have converted.
• Presenting a single number rather than a range across assumptions.
• Counting repeat attempts from the same caller as separate lost opportunities.
• Ignoring the calls that were answered and then went nowhere, which is often a larger problem than the missed ones.
• Building the case around the most expensive remedy rather than around what the pattern actually shows.
• Using the mean deal value where a handful of large contracts distort it.

What the number cannot tell you

Be explicit about the limits of this analysis when presenting it.

• It cannot tell you which specific opportunities were lost, only the aggregate shape.
• It cannot establish causation for callers who would have gone elsewhere regardless.
• It does not measure the quality of the calls that were answered, which may be the larger problem.
• It says nothing about whether a given remedy will work — that requires a measured trial with a baseline.
• It cannot value the retention and reputation effects, only note them.

Stating these limits in the presentation is what makes the rest of it credible. An analysis that claims more than it can support gets discounted entirely, including the parts that were sound.

Decision framework and next step

Four questions to work through.

1. 1

   Can you separate the six categories in your telephony reporting?

   If not, that is the first task and it is usually a configuration change rather than a project.

2. 2

   Do you know your own conversion rate on answered inbound calls?

   Without it the calculation relies on borrowed numbers and will be dismissed.

3. 3

   What does the hourly distribution look like?

   This decides which intervention is appropriate more than the total does.

4. 4

   What proportion of answered calls did not need a person?

   Sample them. This number determines whether automation returns capacity or simply adds a layer.

Produce the range, fix the cheap items first, and trial one intervention against a measured baseline before committing further. For the wider comparison of operating models, see AI and traditional call centres; the outbound side is covered in automating sales calls. Our AI solutions overview sets out how these programmes are staged.

Frequently asked questions

1. 1

   How do you calculate the cost of missed calls?

   Multiply missed calls from genuine prospects by your own conversion rate on answered inbound calls, by your own average or median deal value, by a conservative recovery assumption. Present the result as a range across several assumptions rather than a single figure.

2. 2

   What data do you need?

   Telephony reporting separating unanswered, abandoned, busy, out-of-hours, unreturned voicemails and answered-but-dropped calls by hour; a classified sample of answered calls; CRM data linking calls to outcomes; and your own deal value.

3. 3

   Why not use industry benchmarks?

   Because they invite the objection that your business is different, and they are frequently wrong for a specific call mix. A figure derived from your own telephony and CRM data is harder to dismiss and more likely to be acted on.

4. 4

   What are the most common mistakes?

   Assuming every missed call was a prospect, assuming every prospect would have converted, presenting one number instead of a range, counting repeat attempts as separate losses, and ignoring answered calls that went nowhere.

5. 5

   What should be fixed first?

   Line capacity problems and unreturned voicemails. Both are usually cheap, both show up clearly in the category breakdown, and fixing them first makes the rest of the analysis credible.

The purpose of the exercise is not to justify a purchase. It is to find out which part of the problem is capacity, which is process, and which — if any — is worth solving with technology.