Read on Twitter
[1/13]
Very excited to share a new paper with Elliot Oblander, rockstar PhD student at @Columbia_Biz. We spent much of the past year studying how, how much, and why COVID-19 impacted restaurant delivery.
Paper: https://bit.ly/330SUSJ
Summary:https://bit.ly/330SUSJ&q... href=" https://bit.ly/2QIF1pV ">https://bit.ly/2QIF1pV&q...
Very excited to share a new paper with Elliot Oblander, rockstar PhD student at @Columbia_Biz. We spent much of the past year studying how, how much, and why COVID-19 impacted restaurant delivery.
Paper: https://bit.ly/330SUSJ
Summary:
[2/13]
The prospects (and for many, survival) of restaurants and restaurant delivery companies hinge on what consumer demand will look like after the pandemic ends. Given the recovery is still a work in progress, we hope this provides some early insight.
The prospects (and for many, survival) of restaurants and restaurant delivery companies hinge on what consumer demand will look like after the pandemic ends. Given the recovery is still a work in progress, we hope this provides some early insight.
[3/13]
A unique collection of data sources allowed us to answer this question.
a)Main source: @earnestresearch, which provided us with transactional data for ~2M panel members from 1/2016-12/2020, plus modal location each month. 99.8% correlation to "gold standard" SEC data.
![[3/13]A unique collection of data sources allowed us to answer this question. a)Main source: @earnestresearch, which provided us with transactional data for ~2M panel members from 1/2016-12/2020, plus modal location each month. 99.8% correlation to "gold standard" SEC data.](https://pbs.twimg.com/media/E0FqEtIWEAEUs64.jpg "[3/13]A unique collection of data sources allowed us to answer this question. a)Main source: @earnestresearch, which provided us with transactional data for ~2M panel members from 1/2016-12/2020, plus modal location each month. 99.8% correlation to \"gold standard\" SEC data.")
![[3/13]A unique collection of data sources allowed us to answer this question. a)Main source: @earnestresearch, which provided us with transactional data for ~2M panel members from 1/2016-12/2020, plus modal location each month. 99.8% correlation to "gold standard" SEC data.](https://pbs.twimg.com/media/E0FqEtLWYAc7mko.png "[3/13]A unique collection of data sources allowed us to answer this question. a)Main source: @earnestresearch, which provided us with transactional data for ~2M panel members from 1/2016-12/2020, plus modal location each month. 99.8% correlation to \"gold standard\" SEC data.")
A unique collection of data sources allowed us to answer this question.
a)Main source: @earnestresearch, which provided us with transactional data for ~2M panel members from 1/2016-12/2020, plus modal location each month. 99.8% correlation to "gold standard" SEC data.
[4/13]
b) @yipitdata: every restaurant listed on every major delivery platform, monthly from 10/2018 to 12/2020.
c) @waybackmachine: historical versions of major delivery company websites, giving us their historical market coverage (and thus entry).
![[4/13]b) @yipitdata: every restaurant listed on every major delivery platform, monthly from 10/2018 to 12/2020.c) @waybackmachine: historical versions of major delivery company websites, giving us their historical market coverage (and thus entry).](https://pbs.twimg.com/media/E0FrDTbXIAkeJSb.png "[4/13]b) @yipitdata: every restaurant listed on every major delivery platform, monthly from 10/2018 to 12/2020.c) @waybackmachine: historical versions of major delivery company websites, giving us their historical market coverage (and thus entry).")
![[4/13]b) @yipitdata: every restaurant listed on every major delivery platform, monthly from 10/2018 to 12/2020.c) @waybackmachine: historical versions of major delivery company websites, giving us their historical market coverage (and thus entry).](https://pbs.twimg.com/media/E0FrDUfWQAQ4vta.png "[4/13]b) @yipitdata: every restaurant listed on every major delivery platform, monthly from 10/2018 to 12/2020.c) @waybackmachine: historical versions of major delivery company websites, giving us their historical market coverage (and thus entry).")
![[4/13]b) @yipitdata: every restaurant listed on every major delivery platform, monthly from 10/2018 to 12/2020.c) @waybackmachine: historical versions of major delivery company websites, giving us their historical market coverage (and thus entry).](https://pbs.twimg.com/media/E0FrDVYWEAYkweT.png "[4/13]b) @yipitdata: every restaurant listed on every major delivery platform, monthly from 10/2018 to 12/2020.c) @waybackmachine: historical versions of major delivery company websites, giving us their historical market coverage (and thus entry).")
b) @yipitdata: every restaurant listed on every major delivery platform, monthly from 10/2018 to 12/2020.
c) @waybackmachine: historical versions of major delivery company websites, giving us their historical market coverage (and thus entry).
[5/13]
d) @SafeGraph: location data for 18M devices, giving us measures of stay-at-home behavior, dine-in activity, and restaurant employment each day by census block group.
[and SEC data]
![[5/13]d) @SafeGraph: location data for 18M devices, giving us measures of stay-at-home behavior, dine-in activity, and restaurant employment each day by census block group.[and SEC data]](https://pbs.twimg.com/media/E0FrKbwXsA0ngZK.png "[5/13]d) @SafeGraph: location data for 18M devices, giving us measures of stay-at-home behavior, dine-in activity, and restaurant employment each day by census block group.[and SEC data]")
d) @SafeGraph: location data for 18M devices, giving us measures of stay-at-home behavior, dine-in activity, and restaurant employment each day by census block group.
[and SEC data]
[6/13]
We use a multi-method approach: an event study and FE regression models to estimate absolute impact and the mechanisms driving it. Unlike prior work, we focus on revenue and how it breaks down into acquisition, retention, ordering and spend.
We use a multi-method approach: an event study and FE regression models to estimate absolute impact and the mechanisms driving it. Unlike prior work, we focus on revenue and how it breaks down into acquisition, retention, ordering and spend.
[7/13]
Main finding 1:
COVID made the year. Industry-wide delivery sales went from ~$23B in 2019 to ~$51B in 2020, or growth of ~$28B. About $19B of that $28B (70% of total) was due to the pandemic. If not for the pandemic, sales growth would have fallen by over half.
![[7/13]Main finding 1: COVID made the year. Industry-wide delivery sales went from ~$23B in 2019 to ~$51B in 2020, or growth of ~$28B. About $19B of that $28B (70% of total) was due to the pandemic. If not for the pandemic, sales growth would have fallen by over half.](https://pbs.twimg.com/media/E0Fub8KWEAc_tpc.png "[7/13]Main finding 1: COVID made the year. Industry-wide delivery sales went from ~$23B in 2019 to ~$51B in 2020, or growth of ~$28B. About $19B of that $28B (70% of total) was due to the pandemic. If not for the pandemic, sales growth would have fallen by over half.")
![[7/13]Main finding 1: COVID made the year. Industry-wide delivery sales went from ~$23B in 2019 to ~$51B in 2020, or growth of ~$28B. About $19B of that $28B (70% of total) was due to the pandemic. If not for the pandemic, sales growth would have fallen by over half.](https://pbs.twimg.com/media/E0Fub91XoAEz1E-.jpg "[7/13]Main finding 1: COVID made the year. Industry-wide delivery sales went from ~$23B in 2019 to ~$51B in 2020, or growth of ~$28B. About $19B of that $28B (70% of total) was due to the pandemic. If not for the pandemic, sales growth would have fallen by over half.")
![[7/13]Main finding 1: COVID made the year. Industry-wide delivery sales went from ~$23B in 2019 to ~$51B in 2020, or growth of ~$28B. About $19B of that $28B (70% of total) was due to the pandemic. If not for the pandemic, sales growth would have fallen by over half.](https://pbs.twimg.com/media/E0Fub-3XoAMX2mh.png "[7/13]Main finding 1: COVID made the year. Industry-wide delivery sales went from ~$23B in 2019 to ~$51B in 2020, or growth of ~$28B. About $19B of that $28B (70% of total) was due to the pandemic. If not for the pandemic, sales growth would have fallen by over half.")
Main finding 1:
COVID made the year. Industry-wide delivery sales went from ~$23B in 2019 to ~$51B in 2020, or growth of ~$28B. About $19B of that $28B (70% of total) was due to the pandemic. If not for the pandemic, sales growth would have fallen by over half.
[8/13]
Main finding 2:
Growth was due to people replacing restaurant visits with delivery orders, either because on-prem dining wasn& #39;t allowed because of dine-in restrictions, or fear of going on-prem for health reasons. All other mechanisms depressed delivery, all else equal.
![[8/13]Main finding 2: Growth was due to people replacing restaurant visits with delivery orders, either because on-prem dining wasn& #39;t allowed because of dine-in restrictions, or fear of going on-prem for health reasons. All other mechanisms depressed delivery, all else equal.](https://pbs.twimg.com/media/E0Fuf0YXEAQ9EWi.png "[8/13]Main finding 2: Growth was due to people replacing restaurant visits with delivery orders, either because on-prem dining wasn& #39;t allowed because of dine-in restrictions, or fear of going on-prem for health reasons. All other mechanisms depressed delivery, all else equal.")
![[8/13]Main finding 2: Growth was due to people replacing restaurant visits with delivery orders, either because on-prem dining wasn& #39;t allowed because of dine-in restrictions, or fear of going on-prem for health reasons. All other mechanisms depressed delivery, all else equal.](https://pbs.twimg.com/media/E0Fuf1YXsAAsYpD.png "[8/13]Main finding 2: Growth was due to people replacing restaurant visits with delivery orders, either because on-prem dining wasn& #39;t allowed because of dine-in restrictions, or fear of going on-prem for health reasons. All other mechanisms depressed delivery, all else equal.")
![[8/13]Main finding 2: Growth was due to people replacing restaurant visits with delivery orders, either because on-prem dining wasn& #39;t allowed because of dine-in restrictions, or fear of going on-prem for health reasons. All other mechanisms depressed delivery, all else equal.](https://pbs.twimg.com/media/E0Fuf2WWEAEX4_z.png "[8/13]Main finding 2: Growth was due to people replacing restaurant visits with delivery orders, either because on-prem dining wasn& #39;t allowed because of dine-in restrictions, or fear of going on-prem for health reasons. All other mechanisms depressed delivery, all else equal.")
Main finding 2:
Growth was due to people replacing restaurant visits with delivery orders, either because on-prem dining wasn& #39;t allowed because of dine-in restrictions, or fear of going on-prem for health reasons. All other mechanisms depressed delivery, all else equal.
[9/13]
Main finding 3/implication:
Dine-in activity returning to pre-pandemic levels could spur on a significant headwind for delivery sales. That& #39;s how substitution works. If growth was due to delivery orders replacing restaurant visits, then if restaurant visits roar back, ...
Main finding 3/implication:
Dine-in activity returning to pre-pandemic levels could spur on a significant headwind for delivery sales. That& #39;s how substitution works. If growth was due to delivery orders replacing restaurant visits, then if restaurant visits roar back, ...
[10/13]
A lot more in the paper. If you& #39;re crunched for time, I would recommend the summary: https://bit.ly/3vashrn ">https://bit.ly/3vashrn&q...
A lot more in the paper. If you& #39;re crunched for time, I would recommend the summary: https://bit.ly/3vashrn ">https://bit.ly/3vashrn&q...
[11/13]
Two caveats given what I usually write about:
a) Dine-in won& #39;t come back overnight. Dine-in regulations will be an important first step, but there are also vaccines and more importantly, changes in restaurant operations -- waitstaff and configuration of restaurants.
Two caveats given what I usually write about:
a) Dine-in won& #39;t come back overnight. Dine-in regulations will be an important first step, but there are also vaccines and more importantly, changes in restaurant operations -- waitstaff and configuration of restaurants.
[12/13]
b) This isn& #39;t $DASH bashing. I was constructive on $DASH economics, even pre-pandemic: https://bit.ly/3lqUAfk
I">https://bit.ly/3lqUAfk&q... still believe it!
$DASH is also outperforming the category. They note it themselves here: https://bit.ly/3gIKcRn ">https://bit.ly/3gIKcRn&q...
b) This isn& #39;t $DASH bashing. I was constructive on $DASH economics, even pre-pandemic: https://bit.ly/3lqUAfk
I">https://bit.ly/3lqUAfk&q... still believe it!
$DASH is also outperforming the category. They note it themselves here: https://bit.ly/3gIKcRn ">https://bit.ly/3gIKcRn&q...
