Eric Anderson and Duncan Simester published a paper* on pricing in 2003 that showed that $9 price endings increased demand in three experimental settings. The benefit was greatest when the products were new and not previously sold.
They were less effective on sale items. And they should not be used on every item.
This obviously predates the prevalence of round pound and price-marked packs in the UK convenience channel.
I think the point to note is that you need to mix it up.
Earlier this year I suggested to one retailer that he did not have any promotions in his shop.
He said it was because it was so hard to change prices on his IT system. My reply was to just keep the prices the same but to promote one SKU harder where his shoppers would expect to see a promotion. I have not yet been back to see what he did.
This has no scientific evidence to support it. But I have appended it to a report about a scientific study so perhaps you will be persuaded to act on it like you were a shopper approaching a promotional bay in store!
*Effects of $9 Price Ending on Retail Sales: Evidence from Field Experiments