Knowles Knowledge Centre•Diagnosis

How to Identify Whether Email Is Costing Your Team More Than It Saves

A diagnostic method for working out whether an unmanageable inbox is a writing problem, a copying expectation or unclear work — before you commission any training.

An email skills gap exists when senders cannot produce a message that costs less than it achieves, even where they are free to shorten a recipient list, the work has clear owners, and the team has agreed what gets copied. That is a narrower condition than it sounds. Most unmanageable inboxes are produced by an expectation that senior people are copied, by work whose ownership nobody has settled, or by the absence of a single sentence stating what happens if nobody replies.

This guide sets out a three-step diagnostic method, the evidence sources that support it, sample questions for separating a writing gap from an expectation, and thresholds for deciding when development is the right response.

Practical guide•For HR and L&D teams•Updated September 2026

About this guide

Published byKnowles Training Institute
AudienceManagers and anybody about to commission business writing training
Use beforeResponding to a complaint that email is out of control

Why email volume is misdiagnosed

The common failure is starting from the inbox. People report that email is unmanageable, and the finding enters the pipeline as a writing problem rather than as a question about how many names are on each message and why they are there.

Three things go wrong when the complaint is accepted at face value.

A skill is inferred from a volume“Our email is out of control” describes a quantity and names no behaviour. Nine-name recipient lists, replies that are only confirmations, and threads nobody closes produce the same inbox and require three unrelated remedies, and the complaint cannot distinguish them.
Writing quality is mistaken for costA well-written message is convincing to everybody including its reader. The subject line is clear, the ask is single, the tone is right. Nobody has counted how many people opened it or how many replies it generated, which is the only measurement that separates an efficient message from a pleasant one.
No alternative explanation is testedIf a writing course is the only option considered, it will be the option chosen — whether senders were free to shorten a list, whether the work had owners, or whether anybody had ever stated a default.

A badly written message and a well-written message to nine people produce the same inbox. Only one of them responds to training.

The count–list–cause method

The method works backwards from the inbox to what was actually sent, then tests whether sender behaviour is the cause.

1

Count what one team sent

Take a fortnight of one team’s sent messages and count. “Median nine recipients, two replies per message, each reply to all nine” is usable. “We get too much email’ is not.

If the complaint cannot be stated as a count of recipients and replies, the diagnosis stops here until it can, because a volume with no structure in it has no cause and no remedy.

2

Test the list against the question

For each message, ask what each recipient’s answer would have changed. Sort the names into three groups: needed for an answer, being informed, and copied as insurance. The third group is the finding, and it is usually larger than anybody expects and readily admitted once asked.

For one team the median message may need two answers, inform three people and copy four as insurance. The four are not a writing failure; they are a response to an expectation somebody set, and they will not be removed by anything a sender learns.

This is the step most organisations skip. Without it, everything downstream is an assessment of senders.

3

Test whether sender habit is the cause

Ask the question that separates a habit from an expectation: is a sender free to leave a senior name off a message?

If not, the copying is an instruction rather than a habit, and no amount of practice reaches it. The remedy belongs to whoever set the expectation, and until it changes the recipient counts will not.

If they are free, establish whether requests carry dates, whether any message states a default, and whether anybody closes threads. Those three point at planning, at one missing sentence and at an unassigned job, and the remedies differ and are all cheap.

Where the evidence comes from

A diagnosis is only as good as what it is built on. Most of this can be settled by counting a fortnight of sent messages, and organisations reach instead for the source that measures how the inbox feels rather than what arrives.

Evidence source

What it shows well

What it cannot show

A fortnight of one team’s sent messages

Median recipients, replies generated and who they reach

Why each name was on the line

The senders, asked why each name was there

Which names are answers, which are information, which are insurance

Whether the insurance was ever required of them

Whoever set the copying expectation

Whether a sender may leave a senior name off

Whether anybody believes the answer

A sample of requests, read for deadlines

Whether an urgency or a date was given, and its reason

When the thing was actually needed

A sample of threads, read for closings

How many ended with a stated outcome

Who everybody thought was going to close them

An inbox satisfaction survey

How burdensome people find their email

Anything about what was sent — a nine-name habit and a heavy workload feel identical

Note the pattern: the source most organisations lead with — how people feel about their inbox — is the only one on this list that cannot distinguish a sending problem from a workload problem, which is the first thing any remedy has to establish.

Use at least two independent sources before concluding that writing is the gap. A fortnight’s counts supported by the senders’ own account of why each name was there is evidence. A satisfaction score is a hypothesis.

Diagnostic questions that separate cause from symptom

Ask these in sequence. The first “no” usually identifies the constraint.

FreedomIs a sender free to leave a senior name off a message? Where copying upwards is read as transparency and omitting it as concealment, the list is not the sender’s to shorten.
OwnershipDoes the work have clear owners? Where nobody knows who decides, messages go to everybody who might, and the volume is a symptom of the ambiguity.
DatesDo requests carry a date and its reason? An urgency asks every recipient to guess, and their guesses will not agree, which generates clarifying replies.
DefaultsDoes any message state what happens without a reply? Where silence has no consequence, recipients reply to confirm, and those confirmations are a large share of the traffic.
ClosingIs closing a thread anybody’s job? Where it is nobody’s, threads stay open in every inbox on them, and the cost is paid daily by people with nothing left to do.
CapabilityWith senders free to shorten a list, work that has owners, dates on requests, defaults stated and somebody closing threads — is the volume still a problem?

Only the last question identifies a development need. The first five identify an expectation, an ownership problem and three missing habits that cost nothing to adopt — and none of them is addressed by a writing course.

Decision thresholds

A diagnosis should end in a decision, not a report. These thresholds keep that decision defensible.

Finding

Threshold

Response

Senders cannot shorten the list

Copying upwards is expected, and omitting it is noticed

An expectation problem. It belongs to whoever set it, not to senders.

Nobody knows who decides

Messages go to everybody who might own the matter

An ownership problem. The traffic is a symptom of the ambiguity.

No request carries a date

Urgencies throughout, across senders

A habit, and the cheapest to change. It needs a sentence, not a course.

No message states a default

Most replies are confirmations rather than answers

A habit, and the one that removes the most traffic.

Free, owned, dated, defaulted, closed

One sender’s counts are far above comparable colleagues’

Now it is individual, and development is appropriate. Scope it to the stage.

The threshold that matters most

If senders are not free to leave a senior name off a message, there is no evidence about anybody’s email skill and no basis for training them. That single check prevents more pointless business writing spend than any other step in this method.

A worked example

A Singapore insurance firm’s operations team reports that email is unmanageable. A mailbox hygiene campaign is run, and six months later the volume is unchanged. Business writing training for fourteen people is now proposed.

1

Count what was sent

A fortnight of the team’s sent messages is counted. The median message reaches nine people and produces two replies, each going to all nine. Nobody in the team had ever seen the aggregate, and the heaviest senders are as surprised as anybody.

2

Test the list

Senders are asked what each name was for. A median of two answers were needed. About three recipients per message were being informed and could have had a single later note. The remaining four were copied as insurance, which five of the fourteen volunteer without being prompted.

3

Find the cause

The expectation is traced. Two years earlier, after a complaint reached a regulator, a director asked to be copied on anything client-facing. The request was reasonable, was never bounded, and has been read ever since as a general instruction. Separately, two thirds of the requests carry an urgency rather than a date, and almost no message states a default.

Diagnosis: this is not mainly a writing gap and a course for fourteen people would have been aimed at the wrong thing. Four names per message are on the line because of an unbounded request nobody has revisited, and a large share of the replies are confirmations produced by the absence of one sentence.

Response: a business writing course would have improved messages that should not have reached nine people. What is needed is the director’s request bounded to a defined set of matters, an agreement in the team about dates and defaults, and somebody named to close each thread. One narrower need is real: two of the fourteen send at counts well above their colleagues on comparable work.

Frequently asked questions

?
What is an email skills gap?It exists when a sender cannot produce a message costing less than it achieves, despite being free to shorten the list, the work having owners, and the team having agreed what gets copied. If any of those is absent, what you have is an expectation, an ownership or an agreement problem rather than a writing one.
?
How do you measure it objectively?By counting a fortnight of one team’s sent messages: median recipients, replies generated, who those replies reach, and the proportion of requests carrying a date. That takes an afternoon and produces numbers. Satisfaction scores cannot do it, because a heavy workload and a nine-name habit feel the same from the inside.
?
Should we audit everybody’s mailbox?No, and it would answer the wrong question. What arrives in an inbox is produced by other people’s sending, so the useful measurement is of what one team sends. Auditing what people receive produces a league table of victims and no remedy.
?
What if people insist the volume is the problem?They are usually right that it is unmanageable and wrong about the cause. The disagreement is settled by the counts rather than by discussion: a median of nine recipients where two answers were needed is a finding nobody disputes, and it moves the conversation from habits to the expectation that produced them.
?
How often should this be repeated?Tie it to the work rather than the calendar. Re-count after any change to who is copied by default, because an unbounded request from somebody senior is the commonest single cause; after any reorganisation, since unclear ownership sends messages to everybody who might decide; and a month after any agreement about defaults, because those decay quietly. Annual surveys tend to produce ritual rather than insight.

Conclusion

Identifying an email problem is mostly an exercise in counting what was sent. The method is not complicated:

1Count a fortnight of one team’s sent messages

2Sort each recipient into answer, information or insurance

3Ask whether a sender may leave a senior name off

The discipline it requires is the willingness to conclude that nobody’s writing is the problem.

A diagnosis that recommends training every time it is run has not been a diagnosis; it has been a procurement process.

Where the diagnosis does point to sender behaviour rather than to an expectation or the work, the next question is which stage — covered in what an email costs the person who opens it. Organisations that decide facilitated development is appropriate can review the Writing Effective Emails Training Course in Singapore, evaluating scope and expected workplace application against the diagnosis rather than against the course description.