Designing Apps for Viral Growth

Viruses are notorious for how quickly they can become prevalent in a population. So are viral apps. Can epidemiology help us design more viral apps?

We know from disease control that any virus that has an Ro > 1 will spread exponentially throughout a susceptible population. Many attempts to slow the spread of a viral disease center around interventions to reduce its effective R0 to be less than 1. If something can sustain an effective R0 < 1, it will decline in prevalence and eventually go away completely. This is the reasoning behind public health measures such as wearing masks, social distancing, checking temperatures, pursuing early detection, and racing for a vaccine when dealing with a viral respiratory disease such as COVID-19 and its viral agent SARS-Cov-2.

When we think of epidemics, we generally think of the ones that make us sick. It is true that epidemiology is primarily concerned with disease control. However, the Greek definition of Epidemiology is “the study of that which befalls people.” Although most applications of epidemiology are in public health and disease control, epidemiology is in fact a general science. Its statistical and investigative methods can be applied to measure and define the spread of anything that befalls a population. We can use epidemiology to measure the spread of memes, ideas, and even businesses, because like viruses and diseases, memes, ideas, and businesses can also befall a population.

The most popular apps and web services spread like a virus in their early stages. The facebook epidemic (and now facebook pandemic) had an estimated  R0 of 8. On average, every new facebook user brought eight additional people to facebook with them. The Dropbox epidemic had an estimated R0 ≈ 2. Every new Dropbox user brought two additional people with them. The Hotmail epidemic also had a formidable R0. So formidable, that Microsoft’s email offering was never able to catch up with Hotmail’s user base, and Microsoft had no choice but to buy Hotmail. And although not as well-known in the US, no social network could match the Bebo epidemic‘s growth in the marketplaces where it had a foothold. The Uber epidemic certainly had an R0 > 1. And everybody remembers the Am I Hot or Not epidemic. You may find it strange that I am referring to internet companies as epidemics. We generally don’t observe of the exponential growth of apps as they quickly become prevalent and call them epidemics. But they share the defining characteristic of the things that we do consider to be epidemics – the rapid spread from few people to a large number of people in a given population, spread from person to person.

While we often talk about digital media going viral when they do, we rarely analyze their virality from an epidemiological lens. How many epidemiologists focus their research on memes? Most have much more meaningful and purposeful fish to fry, and for very good reason. Few people go through medical school to study the spread of memes when they have the brains to save millions of lives in the public health sector.

So while I do not claim to be an epidemiologist, this is an academic interest of mine, and I have more than just a passing interest in the various interdisciplinary topics. I think I may be able to add to this conversation by distilling my personal thoughts and observations onto this humble page. Please note that these thoughts are my own and are not endorsed by any institution.

How to Design Viral Apps

Or, the things you should do to help make sure you’re not crippling the potential exponential spread of your own app.

When a novel and deadly virus is discovered spreading through a population, the first thing that epidemiologists try to ascertain is its R0. Viruses mutate all the time, and each strain can have a different R0. The vast majority of new viruses have an R0 < 1 and will not become epidemics. Although viruses are cataloged as they are discovered, the general public rarely ever hears about the ones that have an R0 < 1, since they will likely not become public health concerns. However, every once in a while, a new virus is actually virulent. We start hearing about more and more people getting sick from an unknown disease. If its R0 is ascertained to be > 1, health experts will sound the alarm bells, and we take the threat seriously. We try to do what we can to reduce its effective R0 to be < 1. In practice, this is accomplished by weakening or breaking the linkages in its epidemiological triad of hosts, agents, and environment.

However, when a new app is created, the last thing that the app developers do is try to ascertain its R0. Like most viruses, the vast majority of new apps have an R0 < 1. So an app developer is highly unlikely to release a viral app by chance. However, every once in a while, a new app spreads virally through a population. So while public health workers work to reduce the effective R0 of a disease, the goal of an app developer should be to mutate their app, with the goal of creating or strengthening linkages in the epidemiological triad of hosts, agents, and environment. To tweak the app iteration after iteration until they have released a strain with an R0 > 1. At that point, it will spread virally.

It is my opinion that the number one thing that an app can do to ensure its success is to be viral. However, most apps forget to be communicable. And if their app’s fundamentals make it so that there can’t ever be a conceivable strain with an R0 > 1, then they should probably work on a fundamentally different app if their goal is to produce a viral app.

But how does an app developer do that? What levers can they pull? This article focuses on this topic from an epidemiological and managerial lens. It’s essentially the study of epidemiology as applied to an app business. As a living document, its aim is to define terminology, mechanisms, and formulas that can be used to measure and influence the virality of apps.

This article will be updated with new sections from time to time.

Chadwick Boseman

An hour ago, Barack Obama shared this post:

Chadwick came to the White House to work with kids when he was playing Jackie Robinson. You could tell right away that he was blessed. To be young, gifted, and Black; to use that power to give them heroes to look up to; to do it all while in pain – what a use of his years.

“Chadwick came to the White House to work with kids when he was playing Jackie Robinson. You could tell right away that he was blessed. To be young, gifted, and Black; to use that power to give them heroes to look up to; to do it all while in pain – what a use of his years.” – Barack Obama

“To do it all while in pain.” That’s the part that gets me the most. That he knew his life might be cut short (or that it might be just long enough) and decided that this is how he would live it. That he would use his talents to depict heroes and fight battles on camera, while few people knew that he was fighting an even bigger battle off camera.

To do it all while in pain, Chadwick was a man living a life with purpose. He knew that he was called to temporarily occupy this time and place in history for a purpose. He did not strike many of us as a man whose time would be ending soon, although he probably knew this is the way his story might end while undergoing his treatments between filming.

To do it all while in pain, his graduation speeches, public appearances, and interviews – they should all be given another look under this new lens that few people had. There may be more that he left for us to see beyond his ground-breaking accomplishments on the silver screen.

What are our COVID-19 Key Performance Indicators?

What are our COVID-19 Key Performance Indicators? As a *general* rule, a good KPI is one that is not only meaningful, but can go both up and down. If it can only go one direction, then it is likely to be what is known as a vanity metric. An example is the cumulative death count graph. It can only go up. We generally want to avoid using vanity metrics to inform decision-making, since they’re not good at providing feedback on the impact of our decisions. Instead, here is one graph I like to track. It shows the US’s daily new deaths. Each day, we can have more new deaths or less new deaths than the previous day. It grows exponentially when we do nothing. So we can use this graph to assess whether our efforts have been effective. Fewer deaths than the previous day are good. More deaths than the previous day are bad. Weaker exponential growth would also be a good sign, but it’s not as easy to spot.

It has a two week lag, so while the media is sounding the alarm bells over 16,000+ cumulative deaths (and rightfully so), many of these deaths were unfortunately locked in weeks ago due to our activities weeks ago. However, in the past two days, the daily number of new deaths has gone down day after day. So that’s a good sign, if this trend keeps up. Maybe I should wait a few more days to see if this trend continues.

One thing to note though is that due to this being a graph showing all US deaths, it’s possible that this is just the result of a state with many cases like New York clamping down hard. And so states with orders of magnitude fewer cases may still be on a steep exponential growth trajectory and this graph wouldn’t show that. Since their numbers are smaller, their exponentially growing numbers wouldn’t move this needle yet. We’d have to see this graph at the state level for each state to assess how individual states are doing.

© 2020 Lawrence Weru Blog

Theme by Anders NorenUp ↑