Google announced yesterday the initial roll out of a new “store visits” metric for AdWords. Essentially, it is an attempt to trace conversion from an ad click to an actual store visit. According to Google:

With the holiday season upon us, it’s clear that the majority of sales for many industries still happen in person – in fact, roughly 95% of retail sales take place in physical stores.¹ And online activities are influencing offline transactions more than ever, bringing together the digital and physical worlds. Thirty-two percent of consumers say that location-based search ads have led them to visit a store or make a purchase, so it’s more important than ever for businesses to understand the impact that search ads have in driving visits to your physical locations, whether that’s a store, hotel, auto dealership or restaurant.

The implementation, however, leaves something to be desired. Google will establish location by conventional means, e.g. geo-fencing and Wi-Fi, and which can have an error rate of over 500 meters!

This means the store visit metric will only work for certain kinds of retailers. It won’t work for stores in dense areas or in shopping malls. It’ll only work for a Costco-like mega store that’s in the middle of nowhere by itself.

Worse, a store visit will count up to 30 days from when an ad is clicked. 30 days! I don’t even remember what I did last weekend. If the ad worked, you’d think the user would have acted on the ad sooner than 30 days.

Yet, Google’s effort is better than nothing and the message is a good one.

The advertising industry is ripe for disruption – even the online version of it. We’ve done our own studies to establish the link between plain vanilla online ads and store visits, and our conclusion is that conversion is not great. To the order of about $100 per paying customer.

Google has its own case study in PetSmart. Supposedly, PetSmart achieved 10% to 18% conversion of search ad clicks to store visits.

Let’s go with 10% as the number, because the definition of a store visit is so generous, the real number is likely far lower. Say you paid $3 per click – a reasonable base cost of a high quality keyword. That means each store visit costs $30, minimum. Not everyone will transact, and the rule of thumb in retail is 20% to 50% will.

Overall, that’s probably less than our own estimate, but then again we have no incentive to make this look good.

While the store visit metric isn’t accurate enough to be definitive, and will no doubt confuse and mislead some, I applaud Google’s attempt to add transparency to the process.

However, when the light is turned on, people may not like what they see.