This is an interesting question! I too was curious, but I haven't been able to find a particularly substantive answer.
It seems self-evident that sites needing to be on "best behaviour" (sorry, I think I mean "having a well-established and effective food-safety culture" ) should be performing better more of the time, but it'd be nice to see some real data.
BRC published the attached White Paper approx. 5 years ago and this seems to support the assumed implications.
IFS have a bit of blurb on their website about it, but it's more marketing than data analysis IMO: https://www.ifs-cert...nnounced-audits
Arguably only tangentially relevant to our industry, but I did also stumble across the attached thesis looking at unannounced audits of safety practices in labs. and this does actually put some numbers (and even a graph or two!) together to support the conclusion that unannounced audits do indeed improve operating practices over time.
Thks for the downloads.
I (so far) looked over Pgs 1-9 of the 1st (Food) document.
Interesting that 84% of the respondents stated their reason for choosing unannounced audit (UA) was "customer requirement".
The survey regretfully makes no mention of characteristics such as -
Types of Production/Risk levels.
Sizes of Staff/Production.
I hypothesise that the majority of respondents possessed >= A Grade. The vast majority of respondent's unannounced audit results appear similar. (I deduce that none of the responding, unannounced, 283 sites (ca.25% of unannounced population) had Nonconformances > Minor NCs. Remarkable ?)
Pg8 summarises the data via 3 Sections - (1) Nonconformities and Grades, (2) Impact of an Unannounced Audit, (3) Comments Received by the Survey
Section (1) contains some conclusions presumably based on data not shown here so no specific comments. Interesting to know that NCs found in announced /UAs are typically similar.
Section (2) summarises data results. Some (Table) terminologies are ambiguous from a FS POV, eg accuracy, thoroughness. (i) I was unable to determine how the quoted overall beneficial value of 50% was calculated (seems high). (ii) The data ( table segments 2,9) appears to not support the use of UAs from a FS POV. (eg 85% responses said UA had no effect on their approach to Food safety). (iii) For unspecified reasons, 58% respondents said they would prefer not to continue with unannounced audits.
Section (3) lists a range of 18 Participant Comments following the Unannounced Survey. 89% (16/18) were IMO "complimentary" . This latter value seems "strangely" high when compared to the datum in (2[iii]) above..
Overall comments regarding UAs under discussion - (a) the sample may not be fully representative, (b) a significant benefit from a FS POV is IMO not yet proven.