Operations
Part of What matters most in corporate communications
Corporate communications benchmarks: baselines, ranges and outliers
Corporate communications benchmarks only work when the audience, period and method match, so compare medians and ranges before any industry average.
What to take away
- A benchmark answers a decision, so fix the population, period and method before you quote any number.
- Report the median and the range; a single average hides the outliers that usually matter most.
- Message pull-through of 60% or better on a tested action is a working floor, not a target to celebrate.
- Correction time after a verified error is the accuracy measure most teams forget to track.
- Compare routine periods with crisis periods separately, never in one league table.
Corporate communications benchmarks fail when unlike audiences, channels or methods land in one league table. Start with the decision the number will change, then fix the population, period and denominator before anyone collects data.
The baseline figures worth tracking
| Measure | Common range | Outlier signal |
|---|---|---|
| Message pull-through | 40% to 70% of tested audience | Below 40% means rewrite and retest |
| Employee engagement | 60% to 75% favorable | Above 85% invites a method audit |
| Share of voice | 15% to 35% of defined coverage | Above 50% usually means a narrow set |
| Media response rate | 5% to 15% of pitched outlets | Above 25% suggests paid placement |
| Correction time | 24 to 72 hours | Past 72 hours, exposure compounds |
| Support resolution | 70% to 90% within target | Below 60% means capacity, not messaging |
| Internal newsletter open rate | 30% to 45% of recipients | Above 65% suggests a small or self-selected list |
| Spokesperson quote pickup | 20% to 40% of placed stories | Above 60% points to a relationship effect |
Treat every figure above as a starting range to test against your own history, not a published standard. Where a number matters and you do not have it, divide the count you observe by the population you defined and record both.
Benchmark ranges and outlier signals
Measure
- Message pull-through
- 40-70% tested audience
- Employee engagement
- 60-75% favorable
- Share of voice
- 15-35% defined coverage
- Media response rate
- 5-15% pitched outlets
- Correction time
- 24-72 hours
- Support resolution
- 70-90% within target
Common range
- Message pull-through
- Below 40%: rewrite, retest
- Employee engagement
- Above 85%: method audit
- Share of voice
- Above 50%: narrow set
- Media response rate
- Above 25%: paid placement
- Correction time
- Past 72 hours: exposure compounds
- Support resolution
- Below 60%: capacity issue
Outlier signal
- Message pull-through
- Employee engagement
- Share of voice
- Media response rate
- Correction time
- Support resolution
Published measurement standards exist, and none of them set a target for your program. AMEC's Barcelona Principles, first agreed in 2010 and revised as version 3.0 in 2020, cover evaluation practice in public relations. Edelman's Trust Barometer has run every January since 2001, surveying trust in business, government, media and NGOs.
The U.S. Office of Personnel Management explains interpreting employee survey results by weighing organizational context and making careful comparisons. That guidance covers the Federal Employee Viewpoint Survey, so its norms do not transfer to a private-company scorecard.
Private teams can check common corporate communications questions on ownership, audiences and approvals.
Where the outliers come from
An outlier is rarely a communications triumph. A share of voice above 50% usually means the coverage set was drawn too narrowly. Engagement above 85% often signals a sampling problem or a survey fielded right after good news.
A working outlier rule: check the data before celebrating any result more than 10 points above the top of its range.
The UK Government Communication Service Evaluation Cycle framework separates inputs, outputs, outtakes, outcomes and impact. It was built for government communication, but the sequence stops a team from treating volume or media attention as the result.
Set a decision rule beside each number. If fewer than a stated share of tested users can name the required action, rewrite and retest. If support wait time rises beyond capacity, slow distribution or add staff. A dashboard without response rules is decoration.
- Keep the denominator visible
- Label surveys, tests, analytics and estimates
- Record every method change
- Separate routine and crisis periods
- Retire measures that never alter a decision
Make the comparison reproducible
The GAO evaluation design guide connects evaluation questions with evidence needs and design choices. Federal evaluation guidance does not make a local marketing result causal or transferable.
The NIST experimental design selection guidance starts design choice with the objective and the practical constraints. It supports keeping benchmark reporting separate from controlled effect estimates.
Before comparing results, state the population, the period, the inclusion rules and the calculation. Add the currency, the maturity window and the uncertainty. Recalculate the figure from the cited source or governed dataset, then record any mismatch. A comparable ladder of definitions keeps each benchmark tied to the decision it informs.
Common questions
What is a good corporate communications benchmark?
It is a defined reference that helps an owner decide something: a prior comparable result, a tested target, a service standard or a qualified external norm. Without the decision attached, the number is trivia.
Should share of voice be a primary measure?
It can describe a defined set of coverage, but it does not prove understanding, trust or favorable behavior. Use it alongside message pull-through and one outcome measure, not instead of them.
How many measures should a small team use?
A few balanced ones tied to current decisions. Delivery, understanding, action and service form a sound starting set. Add an accuracy measure and one relevant outcome, and only when someone will act on the result.
