- A line failed by most sites is a standards problem. A line failed by a few is an execution problem.
- Standards fail for three usual reasons: impossible in the time, unclear, or measuring the wrong thing.
- Fixing a bad standard recovers credibility for every other standard you have.
Read the failures across sites, not down one site
Audit reports are usually read one site at a time, which makes every failure look like a site problem. Reading the same line across every site answers a different and more useful question.
If eighteen of twenty sites fail the same line, you do not have eighteen struggling managers. You have one standard that does not work, and every week you spend coaching against it costs you credibility.
The three reasons standards fail everywhere
It is impossible in the time available, which happens when the standard was written by someone who has not worked the shift. It is unclear, so each site interprets it differently and most interpret it wrongly. Or it measures something that does not matter, so sites deprioritise it correctly.
The three have different fixes: change the resourcing or the scope, rewrite it as a step somebody could follow, or withdraw it.
Read each audit line across every site before reading each site down its lines. Failures that cluster by line are yours to fix, not the sites .
Ask the sites that pass
When most sites fail a line and two pass, the two are the most valuable source you have. Sometimes they have found a genuinely better method worth spreading. Sometimes they are interpreting the standard more loosely than everyone else, which tells you it is unclear.
Either answer moves you forward, and both are faster to obtain than another round of estate-wide coaching.
Withdrawing a standard is a strength
Leaving a broken standard in place because withdrawing it looks like backing down is the expensive option. Every week it survives teaches the estate that standards are aspirational.
Withdrawing it publicly, with the reason, does the opposite: it says the standards that remain are real, which is the thing you actually need people to believe.
Frequently asked questions
What proportion of sites failing means the standard is wrong?
As a rule of thumb, if more than half fail a line consistently, treat it as a standards problem until proven otherwise.
How do I rewrite a standard that is unclear?
Write it as a step a new starter could follow without asking a question, then test that claim on an actual new starter.
Does withdrawing standards undermine authority?
The opposite. Leaving unachievable standards in place is what teaches people that standards are optional.