Thread Time:
Subject: Negative brand impact of PAA results.

Google has been including more balanced questions into People Also Ask (PAA) results. Here is a selection where I think they go to far. This selection was found in 7 searches at random for companies that came to mind.

Many of these are leading and loaded questions which can influence the user's perception of the brand. I should mention that I altered the results to remove Ads and GMB for a cleaner view. 2/20

The issue is that many of these searches are done by a very small fraction of users, yet make visible to all the brands prospects and customers the seed of doubt on some aspect of the company. *You can have 30 searchers influencing brand perception for millions of people.* 3/20

We know that the results are not driven purely by search interest based on several examples we have for major brands with similar results. In many cases, the negative questions are asked by only a handful of users over a 6-month period. 4/20

Further, in most cases, tens or hundreds of other question intents have much more user interest in search. 5/20

I decided to look more broadly at this by using @nozzleio data and the Senta tool from @BaiduResearch. I obtained a list of Fortune 500 companies from 2019 from Github. 6/20

Adding to Nozzle and pulling the PAA results for each company brand search, we found that 85% of these companies had PAA results for Desktop US searches. 7/20

Using the roberta_skep_large_en algorithm, we see it found 63% of the PAA questions for companies to be overall negative in sentiment. 8/20

Using the ernie_2.0_skep_large_en algorithm, we see it found 80% of the PAA questions for companies to be overall negative in sentiment. Either way you slice it, the results have the scales tipped to negative. 9/20

Here is a look at two separate brands with arguably the same product. One has a reflection of web content which reinforces, IMHO, an unfair difference between the two products. They are both really not that great for you. 10/20

Here is the revenue and profit (2019) for brands that have PAA results and those that do not. No PAA is smaller just because there are less. 11/20

Here is the AVERAGE revenue and profit (2019) for brands that do and don't have PAA results. Median values were similarly aligned. Really nothing interesting here. 12/20

Here are some selected beauties that appear for various brand searches. Really love the @Apple callout there. 13/20

For those of you managing brands that have the infamous "Is <company> legal / legit?" as one of the top 4 PAA searches despite only 20 ACTUAL users doing that search, please raise your hand! 14/20

Based on the search volume we see in GSC, this has to be an construct applied to a class of brands and not reflective of true interest in the search for a specific brand. 15/20

Here are the top 20, by revenue, that are free from the scourge of having their brands tarnished by PAA results. They have none. Perhaps we could include:

Does J&J baby powder cause cancer?
Does wells fargo cheat customers?
Is facebook a terrible company?

16/20

Bottom line is that I like the more balanced approach here to these questions. I think that is what Google is going for. 17/20

I really just think the algo is off and causing unnecessary harm for brands that really care about their reputation.
18/20

If I were to get a few thousand people to search "Is Google an evil company?" at some clip for the next 6 months, would Google want that displaying prominently for all billion of their brand searches? 19/20

Of course not, but they really have nothing to fear, since their brand search is coincidentally missing these PAA results. 20/20