Can Google Ads data predict the next coronavirus outbreak?

google news coronavirus map

Do you have the coronavirus? You grab your Phone, head to Google, type “I can’t smell,” and tap the first link that pops up on the page.

What you clicked was a Google Ad. From that one click, Google collects a lot of information about you — demographic data, location, and more.

He isn't promoting a store hawking face masks. Instead, he said he's running Google ads to fight the coronavirus.

Researchers around the world are using search data from Google Trends to track the coronavirus. If there is a sudden searches related to COVID-19 symptoms, it could indicate an outbreak.

But there are lot of a problems with the coronavirus search data Google releases publicly through Google Trends, according to Berlinquette. He says the data is "incomplete" because you can only see "correlations after the fact."

That's why he turned to Google Ads. Once a user clicks on his google ads, the data appears in realtime on a heat map on his website.

Google Trends only provides relative search volume. Berlinquette’s data tells you exactly how many people clicked on his search google ads. He also pointed out that Google Trends doesn’t provide demographic data.

"[Berli's data] surfaces demographic information about the searchers, enabling analysis by age and gender," said Sam Gilbert, a researcher at the Bennett Institute for Public Policy at the University of Cambridge, in an email to Mashable. "This is not possible with Google Trends."

A screenshot of the coronavirus search Google Trends.

Berli’s current project is tracking Google ads clicks in the U.S. related to anosmia, the condition defined by loss of smell, which is believed to be a major symptom of COVID-19.  He just needs them to click on ads so Google can collect their data.

He then displays that data on a public website, Anosmia Google Searches. The data collected from these ads is placed on a map, and broken down in charts by city, gender, and age.

“The idea was that the data would provide epidemiologists, or anyone trying to solve the virus, a new way to find patterns, directly informed by what people are typing into Google,” he said.

So, what does an epidemiologist think of this data? Dr. Alain Labrique, of the Bloomberg School of Public Health and Global mHealth Initiative at Johns Hopkins University, told Mashable that the data could be useful, but too much faith shouldn't be placed in Google searches alone.

He explained how the "gold standard" of data collection is still going into a community and testing to see "what proportion of a population has been infected or is currently infected." Everything else is just "trying to fill in an information gap."

He also said phishing campaigns and scammers looking to take advantage of the pandemic have hindered COVID research.

"It's also very very difficult to figure out how to climb over the mountain of spam to get people to trust who you are and the information you're looking for," he explained.

It's important to note that if a user performs a search on Google, but doesn't click on Berlinquette's ads, they users demographic data isn't recorded for the search marketer. (However, Belinquette notes, that person who doesn't click the ad is still recorded towards the overall search count.)

Dr. Andrew an associate professor of biomedical and health information, He explained how national and local TV news coverage of coronavirus symptoms could affect what people search, and, ultimately, the usefulness of the data.

“Depending on what the president or the governors say, I'm assuming there's a huge spike in search terms anytime they use any one word from vaccine to chloroquine,” Boyd told Mashable. “It's more than just a simple spike in searches.”

“We're talking about a very dynamic situation ... even the fact that you're writing about this article could change people’s search behavior.”

But Berli's tells Mashable that he has planned for that. Before I talked to Boyd, the search marketer asked me to let him know when this piece was published for that very reason.

“I just want to make sure that I’m not dealing with an influx of clicks from people Googling 'I can’t smell' and clicking my ad out of curiosity,” he explained. “I don’t care about the cost, more the dilution of the data. I can do things on my end to prevent it.”

Berli's said that Google Ads data shows him the "word-for-word search" that led to a user clicking his ad. That's why he doesn't run ads on keywords such as “anosmia” or “loss of smell.”

He reasons that someone who finds his ads because they searched “I can’t smell what do I do?” is less likely to have been influenced by a news story than someone who searched "loss of smell." So he runs ads on “I can’t smell,” “lost my sense smell,” and “when you can’t smell.”

A screenshot of Berli's Google search ads.

Mashable reached out to Google with multiple questions regarding this piece. However, the company only replied with information related to its coronavirus-related ad guidelines.

The ads are costing Berlinquette $100 to $200 per day, which he's currently paying for out of his own pocket. Luckily, the search marketer has a full-time job managing Google ad campaigns for 22 businesses.

So, why is Berlinquette doing this? He believes that the data he’s collecting can “predict where infections will resurge once social-distancing rules are relaxed over the coming weeks” and help prioritize where supplies should be shipped.

As for the future of this sort of all data collection, Berli's is looking at the correlation between Google ads and drug abuse and school shootings. He's also involved with a new pilot study at Stanford called Searching for Help: Using Google Adwords for Suicide Prevention.

“It really takes experience in marketing to know how to navigate all the mysterious rules of Google Ads. 
Can Google Ads data predict the next coronavirus outbreak? Can Google Ads data predict the next coronavirus outbreak? Reviewed by Debyendu Bhunia on May 28, 2020 Rating: 5

No comments:

Do not add any spam link in comment box.

Home Ads

Powered by Blogger.