Podcast

Using COVID-19 data to determine business decisions

Marketers and brands are scrambling in response to the “new normal” COVID-19 has created, responding to changing consumer needs by altering the way they do business. At Callahan, we have taken existing COVID-19 data from Johns Hopkins, the World Health Organization and the CDC, overlay it with consumer behavior data, and are identifying how coronavirus will influence consumer behaviors moving forward.

In this episode, Callahan’s VP of data strategy and marketing analytics Zack Pike and chief strategy officer Jan-Eric Anderson discuss how the overlaying of this COVID-19 data will help companies understand the current state of the virus, how it influences consumer behavior and how to use those connections to determine business decisions.

Listen here:


(Subscribe on iTunesStitcherGoogle PlayGoogle PodcastsPocket Casts or your favorite podcast service. You can also ask Alexa or Siri to “play the Uncovering Aha! podcast.”)

Watch the video recording and dashboard demonstration from this podcast conversation

Links to additional COVID-19 podcasts:

What consumers expect from brands during coronavirus
How brands can make responsible coronavirus preparations now

Welcome to Callahan’s Uncovering Aha! podcast. We talk about a range of topics for marketing decision-makers, with a special focus on how to uncover insights in data to drive brand strategy and inspire creativity. Featuring Jan-Eric Anderson and Zack Pike.

Jan-Eric:
Hi, I’m Jan-Eric Anderson, chief strategy officer at Callahan.

Zack:
And I’m Zack Pike, head of data at Callahan.

Jan-Eric:
Zack, it’s great to see you again. Usually we sit a lot closer to each other than we are right now, but we’re holed up at home and trying to make it work in isolation in light of COVID-19, so thanks for finding the time to hop on the interwebs.

Zack:
Absolutely. Yeah.

Jan-Eric:
I know that you’ve been doing some really cool stuff with the team related to COVID-19 data, and I wanted to just catch up with you to see what you guys are doing and how you think it’s useful. So maybe let’s just start with, what are you doing? You’re pulling info in and some data in on what’s happening with the spread. Is that right?

Zack:
Yeah. Yeah. With all the stuff that’s going on, obviously everyone is trying to figure this out. The problem with that is that we’re reliant on all of these medical agencies, the individual counties, states, cities to report their data out. So it took us a little bit of time to find a reliable source. But actually Johns Hopkins has been working on this as well. We were able to hook into one of their data sources that uses the World Health Organization for all of our global data and then the CDC for all of our local stuff. But the cool thing about this data set is also tries to take into account local news reporting because there’s a delay in reporting for several factors and there’s really not a strong standardization in that reporting yet. So that’s the data source that we’re using, which is one that I think a lot of data science teams and analytics people are using right now.

Jan-Eric:
This is reminding me of another conversation I think you and I have had, which is around the vast amounts of free data that is available from third party sources that can be brought in and used for analysis, so yet another example of that. So your intention is not necessarily to try to be the reliable source of reporting for the world. The world’s not going to be coming to Zack Pike to understand what’s going on with the number of coronavirus cases. So what’s the point? Why are you pulling this info together?

Zack:
Yeah, well, so it started … I mean, my team was obviously very curious to dig in because this data provides additional information. I mean, it’s at the lat-long level, so it gives you a lot of geospatial type info, but it also tells you gender and where they think the infection actually originated from for the patient and all kinds of cool stuff.

Zack:
So it was a little bit of curiosity, but also there’s definite application to our clients. So we have clients in the restaurant industry, we have clients in the retail industry in all types of retail and knowing where the virus is, how quickly it’s spreading, when new cases are popping up, can, our assumption is, have big impact on business results in those areas. I mean, we’ve all seen the news on what’s happening with restaurants going to mostly dine out only type of transactions. Different types of concepts are set up differently for that and we’re still digging through the data, but our assumption is that in areas where you see hotspots of this virus activity, you will also see disproportionate impact on sales results. That’s just a hypothesis though. It could prove to be that it’s the same everywhere across the country because everybody’s seeing the same news. We’re all watching the same channels.

Jan-Eric:
So saying another way, as we may be in Kansas City, as we’re seeing information about much higher levels of numbers of cases and subsequent death rates happening, for example, in Washington state, we’re seeing those headlines here in the Midwest, that may be impacting our own behavior even though that’s not necessarily right in our backyard. Is that kind of what you’re saying? There’s a hypothesis that what’s happening locally may be influencing local consumer behavior, but we are all in this together. We may find that actually national news is impacting local behavior as well.

Zack:
Yes. Yeah. You could definitely say that. I think one of our other thoughts with this was if we could start to get a measure on velocity of growth of the virus in certain areas, that may help some companies or some clients get out ahead of what’s eventually going to happen. If you think about a restaurant, if I’ve got restaurants all over the country, some areas are impacted much heavier than others in New York or the Bay Area, I’m already shutting things down because I have to because the government’s making me, but in other areas of the country I’m not. And if I can see OGs in, let’s just say Kansas City, the growth rate of infection is running at the same pace as the Bay Area or New York, I may be able to get ahead of what the government’s going to make me do four or five days, right? And those four or five days could be critical to messaging and adjusting media budgets and just making sure that we navigate this as smoothly as possible.

Jan-Eric:
In your process, generally, I think what you’ll do is you’ll try to get data, relevant data, get access to relevant data, get it structured in a way where you’ve got it within the intelligence platform, the tool that you use, you’ve got it structured in a way that you can start to dig around and play with the data, look at it and through different views to really create some understanding about what’s going on with that data set. Your intention, though, as we’ve been saying, is not necessarily to be the furthermost or the best resource or the most advanced expert on coronavirus, but to understand where it is and how that’s shaping consumer behavior. So have you started to overlay client information or other relevant shopping information or consumer behavior information on top of coronavirus? Have you gotten to that yet? Are you still in the stages of collecting the coronavirus data?

Zack:
Yeah, we just found this data source on Monday, few days ago, and it took us a while to … because of course like any publicly available data source, it’s not clean and it wasn’t in a format or didn’t have the relevant data appended to it that we needed to connect it to that sales performance data. So if you think about sales data, usually the geographic breakdown you have for that at the most minute level is zip code. Well, this data set didn’t include zip code. So we have to take lat- longs, align them to zip codes and do some behind the scenes stuff to get there. But that is the next step is to take the current data that we have and start looking at that up against this virus growth type data.

Jan-Eric:
That’s fascinating. And so then when you start to overlay additional data that you have on top of COVID-19 data, you start to see these patterns and changes in consumer behavior, how some places may be impacted by others, start to have some insight to understanding how does this actually work, what’s the interconnectedness between these different types of activities.

Jan-Eric:
To say that we’re in a state of normalcy right now in the country would be a stretch. In fact, what I keep seeing and hearing more and more is this is our new normal. So there’s probably a bit of a gray area here where we’re going to need to figure out … Pardon my dog while he’s barking. My son’s going to try to quiet him down. What we’re trying to understand is that what is this new normal and how do we adjust in this coronavirus world?

Jan-Eric:
So the intention, the long view of this is really to understand what’s going on. Correct me if I’m wrong, but the intention here is to understand what’s going on with coronavirus, how does that impact behavior, but really how does that influence consumer behavior, but use those connections as a way to steer decisions that marketers are trying to make, that brands are trying to make in this new normal. If we were to make these adjustments, do we think that that’s relevant and what the community needs at large, what consumers are expecting from us? Is it going to match into what their needs are? That’s essentially the long game. Am I right?

Zack:
Yeah. It’s also, I think if we think even longer term, it gives us a framework if something like this ever happens again. So we just got thrown into this. No one was thinking about how you would manage a business through a crisis like this. Usually we’re thinking about PR type crises, big natural disasters, not a virus where largely everyone’s okay, they’re just shut in their house and they’re trying not to do anything. There’s no playbook for something like this. And so there also isn’t on the data side and analytics and so part of this work is trying to navigate our way through it now so that if something like this ever happens in the future, there’s something there that a company can hook onto.

Jan-Eric:
Yeah, that’s great. Well, I think this is really smart in what we’re trying to do and really trying to understand this. Obviously, we’re catching you and talking about this. This is in development. I mean, you’ve just gotten access to the data earlier this week. I know I’ve seen some views, can show kind of what some of the views are of this information that you have, but we’re in the early stages of this. Just walk me through a little bit of what you’ve got here and what we’re looking at.

Zack:
Sure. Yeah. So our focus is confirmed and active cases. We’re not focusing right now on the death rate or really even the recovery rate because those are factors that are probably less influential at this point on consumer behavior unless the death rate shoots way up and we get a lot of them or the recovery rate changes substantially.

Zack:
Our focus, because this is where most of the news media is, is on infection, right? So how many new infections are there every day? Where are those infections? And that’s what this top section here is giving you, right? Now, this data is dated as of about 4:00 yesterday. So there’s actually been a lot that has happened since then. Some of these numbers have actually doubled, but this is showing me by state, where are my hot spots, where are my areas that I have the most infections and then I can drill all the way into county inside the state to figure out where my areas are,, in this case, New York, where in New York we’ve got the most problem, okay?

Zack:
So if you think about if you’ve got locations in a specific area of the country, being able to look at this and say, “Geez, I’ve got a six locations in Harris County in Texas,” and that seems to be one of my strongest areas for infection in Texas and if I haven’t already put a plan in place or haven’t thought as seriously about the plan I should put in place, I really need to be thinking about that at this point.

Jan-Eric:
Yeah. This really does start to give you an idea about how if you’re a retailer or a restaurant brand, how you can overlay then your store-specific data that you have into these geographies and you start to run math to look for the relationship between this data and sales data or search query volume or your ability to drive traffic and foot traffic into a specific location or whatever your sales metric might be.

Zack:
Yep, so that’s raw numbers. But the piece I’m more excited about is growth in infections. And so for anyone looking at this data, it looks like it’s trailing off. That’s not actually true. The way that this data is reported it is it’s reported back to the date that the test was taken, so if you took a test four days ago and you test positive today, you actually add to this 821 here on the 14th. If anybody’s watching this, just keep that in mind.

Zack:
But looking at the growth rate in these areas, especially the areas with lots of infections I think is going to be important, and then putting those growth rates up against each other so we can understand if for some reason the east coast is growing at a different rate than the west coast. You may take a different approach to your response plan based on that growth rate. If the east coast is growing really fast because it’s a much higher density, you may take a different approach than in some areas of the west coast where people are a bit more spread out. They might be taking some stronger precautions with the shelter in place stuff that they already have going in some areas there. You may take a different approach to a retail location or restaurant in that area or different messaging. There’s lots of applications to this type of data.

Jan-Eric:
Yeah. Well, that’s fascinating. And again, the idea here is not necessarily just to be the expert on coronavirus, but to be able to overlay it with other information that gives an indication of consumer behavior and then finding the connections between that consumer behavior and how that might be influenced based on what’s happening with this. This is fascinating. Thanks for pulling that up. Are there any other thoughts that you wanted to mention about this before we wrap up?

Zack:
The only thing I’ll say is, I mean, this is one of the fundamental things we’ve been preaching on this podcast I think since we started, is just making sure you’re leveraging data that’s available to you, right? We didn’t pay any money other than the time we invested to go find this data source and clean it up to be able to leverage it. And yes, World Health Organization, Johns Hopkins, they have trackers out there, but a lot of times they’re not really well-structured for business decisions. Therefore, what they’re designed for, which is to keep track on the virus and what’s happening, but the data’s out there and you just got to take some time and get it structured and use it or work with a company like Callahan who can do it for you, and then this stuff just gives you more arrows in the quiver on making business decisions.

Jan-Eric:
Well, this is fascinating, and Zack, perhaps when you’re a little bit further down the road and have had an opportunity to overlay additional information about consumer behavior, perhaps what we can do is circle back and talk about some of those connections that you’re seeing between the COVID-19 data and other consumer behavior data. This is fascinating. Thanks for coming on and sharing everything.

Zack:
Absolutely.

Jan-Eric:
All right, talk to you again soon.

 

You’ve been listening to the Uncovering Aha! podcast. Callahan provides data savvy strategy and inspired creativity for national consumer brands. Visit us at callahan.agency to learn more.

Watch the video recording and dashboard demonstration from this podcast conversation: