Sam Gilbert writes on search data and the Covid-19 response and how emerging trends could help in the vital search for information on the global pandemic.
“If you work in tech and are wondering what you can do to help with coronavirus”, wrote one Twitter wag in early March, “Wash your hands for 20 seconds, use a tissue for coughs, and avoid touching your face.”
While poking fun at the hubris of the tech industry is fine, this tweet undervalued the role that non-specialist software engineers, data scientists, and even digital marketers can play in response to the crisis.
Here’s one example. One of the many challenges presented by the pandemic is timely access to information in conditions of great uncertainty. Health experts and political leaders need information on which to base their recommendations and decisions; ordinary members of the public need information on what they should and shouldn’t do. This is something that search engine marketers can help with, simply by applying the techniques they would ordinarily use to sell kitchen lighting , slippers, or pet insurance.
How can this be? Every day, people all around the world are turning to Google to find answers to their questions about coronavirus and covid-19. To see this happening in real-time, visit Coronasearch.live, a website built by New York-based search marketer Patrick Berlinquette which shows the searches currently being made by internet users in China. Or look at this visualization by the search marketer Sophie Coley to see how the UK public’s concerns have evolved over the past five weeks from “Should I travel to Italy?” to “Should I close my business?” to “Should I delete Houseparty?” to “Should I shave my head?”
Initiatives like these can stimulate important academic research. A team led by the UCL computer scientist Bill Lampos is working on a model which uses Google searches in Italy for known covid-19 symptoms to track emerging outbreaks elsewhere. It has the potential to predict hotspots in countries with little capacity for testing. One of the model’s most predictive features is the search “perdita olfatto” (Italian for “loss of smell”) which Bill and I had discussed several days before British ENT doctors revealed it was a covid-19 symptom, having both noticed it spiking in Google Trends data.
Meanwhile, data journalist Seth Stephens-Davidowicz has been exploring which symptoms are being searched in unusually high volumes in the US states with the largest numbers of covid-19 cases. This analysis has led him to posit eye pain as an unrecognised indicator of the disease. The timeliness of search data could yet be life-saving.
Search data also offers a vista onto public health information needs. Using Answer The Public, a marketing software tool which captures Google autocomplete suggestions, I have been monitoring the UK public’s top questions about everyday life during the pandemic (my datasets are openly available via the Coronavirus Tech Handbook). This week, the three most-googled questions beginning “Can coronavirus…” are as follows:
1. Can coronavirus live on food?
2. Can coronavirus live on paper?
3. Can coronavirus live on cardboard?
These immediately resonated with me: at home in our flat we have been washing fresh fruit and vegetables in soap and water, and where possible leaving post and deliveries for 72 hours before opening them. This may be out of proportion to the risks or even a complete waste of time, but in the absence of explicit public health advice on these questions it feels like a sensible precaution. My father heats his Sunday newspaper in a low oven before reading it.
Many of the top results on Google searches like these come from websites like The Guardian, where marketers are used to creating and updating online content on a daily basis to win traffic from search engines. This is not something which the teams at gov.uk and nhs.uk would ordinarily need to do, and the last thing they need at this moment of national emergency is more tasks to carry out. So perhaps the answer to closing public health information gaps is for search marketing experts to step up and offer a systematic analysis of these needs, ongoing reporting on how they are evolving, and simple recommendations about what content changes on gov.uk and nhs.uk would drive the most benefit.
Google should help too, by putting more search data into the public domain. Its community mobility reports are a welcome innovation, but if ever there was a time to resurrect tools like Google Correlate and widen access to search data APIs, this is it.
The views and opinions expressed in this post are those of the author(s) and not necessarily those of the Bennett Institute for Public Policy.
