Statistics Canada published an advance estimate before the real retail sales number. Do you trade differently when a preview is already public?

A week before Canada's July retail trade report, Statistics Canada had already published an advance estimate pointing to a 0.8% decline. The real number, released 24 September, came in at a 0.7% fall to $73.7 billion, close to that estimate but not identical (full report).

Statistics Canada is explicit that the advance figure is unofficial and calculated from a partial set of survey responses, roughly 58% of companies in this case, against an average final response rate of 87% over the past year. So it's a real number, but a rough one, and it's already priced into the market by the time the "real" release happens.

That's a genuinely different situation from most releases, where the number is a surprise. We'd like to know how members handle it:

• When an advance estimate is already public, do you treat the "real" release as a non-event unless it surprises relative to the advance figure, rather than relative to economists' forecasts?
• Does the breakdown by subsector or province in the full release ever change your read even when the headline number matches the advance estimate almost exactly, as it did here?
• Do you know which other releases come with an early, unofficial estimate like this, and do you treat them all the same way?
• Has an advance estimate ever been badly wrong, in your experience, and did that change how much you trust the next one?

Please share your own approach and reasoning. Posts that promise a direction or sell signals will be removed.