Corona App: Dead on Arrival

After much back and fro, tinkering, false starts and mind-boggling discussions about data security and privacy German Health Minister Jens Spahn will officially launch the Corona Tracing App in the next couple of days.

It took the combined digital firepower of the country to come up with something that was declared one of the prerequisites to ease the lockdown and reinstall civil rights. The other one being a vaccine.

Cost didn’t matter. Data privacy didn’t matter (at least initially until protests became too vocal and also Google and Apple scuppered the plans for a “centralised solution”). And, efficacy of and sensible reasoning for the app didn’t matter either, it seems.

The background

But, let me throw out some data points first to make sure you understand the context. In a previous article I already described that the draconian lockdown measure the government took were largely unnecessary (see here).

Now that some more time have passed, the picture is getting even clearer. Excess deaths are not getting anywhere to the levels that politician scare mongers were touting.

The run rate for new daily infections has also declined drastically, to around 300…400. According to official figures Covid-19 has claimed 8840 lifes in Germany, the majority of which had been occurring in elderly and social care homes. These places generally appear to be breeding grounds for all different kind of diseases that often go undetected.

The scenario also goes some way towards explaining why the average age of a deceased was 81 years, more than three years above the current life expectancy in Germany. It’s a strong hint as to what should have been done to protect those most vulnerable to Corona.

From randomised field tests (which the government has been ferociously resisting until just recently) we understand that about 10% of the population could well have developed anti-bodies already. This fits with an observation that 80…85% of people are asymptomatic or perceive Corona as just a light flu that doesn’t warrant medical attention.

Too late, too little

Sticking to its line of the App being a defining tool to lift restrictions, Jens Spahn must increasingly be feeling the pushback, and has been backtracking lately. He is now on record saying that he would be happy if the app was downloaded in the millions.

But will this happen. Or are he and his development partners deluding themselves?

Fact is that the case numbers have been petering out for a while now bar the occasional hotspots where people for example have gathered against official advice in restaurants in larger groups, infected themselves on Ramadan parties or failed to keep distance on the latest “activist protests” and lootings (yes, this mindless violence does not only occur in the US).

Other than that, the spread of the virus looks very much under control.

Any takers for the Corona-App, please?

Against that backdrop a consortium of SAP, T-Systems/Telekom, Fraunhofer and the Robert Koch Institute have been developing our new Corona App. To their defence, they were asked to take on this job as an honour. They also published the source code on Github for everybody with coding skills to see and to allay fears of infringements of data security and privacy.

Yet, I believe they will be in for a bit of a disappointment when they finally find out how many people will actually actively use the app. As an app developer you will be trained to go through a (virtual) customer journey before you start designing and coding. That gives you a step-by-step view on sticking points and areas for improvement. 

However, if you were following the announcements for the German Corona App you couldn’t find any of these issues being publicly discussed or addressed. If they were, then it possibly would have become too obvious how hopeless the whole undertaking is.

We know from virologists that, to have a noticeable impact, a tracing app would need to be actively used by ~70% of the population. That’s as in always having your phone with you, switched on and in coverage. Go below this and you simply don’t have the critical statistical spread of the app. The keyword here is “actively used”. It means your phone is ALWAYS on you. Comes rain or shine.

A good benchmark of app proliferation is WhatsApp, the most widely downloaded and installed app in the planet. WhatsApp is running on around 57% of all smartphones worldwide. So at this point already somebody should have stood up and raised the question: “are we on the right track?”

But lets look into some other realities.

85% - ooohh, we are so digital

According to Statista, a mere 85% of the German population has a smartphone. Sounds great. Though, is really everybody using their smartphones? 

80% - ooops, low battery

In reality, people use their phones actively probably only 80% of the time. It’s an optimistic assumption. People forget their phones at home on the dining table, it simmers in the summer heat on the dashboard of their cars or simply runs out of juice.

Or maybe its user has switched off the Bluetooth connection altogether, or the phone fails to get a connection to other phones in its vicinity because it’s buried in that Gucci handbag. It’s also hard to see why people would carry their phones into the public swimming pool or carry it with them when going to play beach volleyball.

Or, are you a football or rugby player and cannot keep your distance? In this case, a remote tackle or socially distanced scrum would be something novel to consider, for sure.

50% - technology, the devil is in the detail

Outright shocking then is the tiny little detail that put technology under the spotlight. The Corona App will only work on Apple iPhones that run iOS13.5. That will count out all the iPhone5s and 6s. Android phone, on the other, can do with OS version 6 to run the app, though many of the latest Huawei phones are banned from the Play Store.

The German smartphone population is rather old (see also here), and I think the developers will be lucky to get the app installed on more than 50% of all the devices out in the field.

80% - now you see me, now you don’t

But it doesn’t stop there. Whoever has worked with communication stacks on smartphones will understand the pain to get two devices to talk to each other. The app uses Bluetooth to exchange data. Bluetooth is fickle. It’s implemented differently by different smartphone manufacturers. It’s configured differently. It may not even want to talk to another phone.

In it’s ~12 years of existence, Bluetooth and Bluetooth Low Energy (BLE) haven’t really cracked that compatibility issue. 

50% - ooouch, how to install a new OS?

A quick look into Statista also tells us that it’s not straight-forward to download and install an app, let alone a new OS on your phone. As of the latest figures (2018) only 50% of user ever installed a new app on their smartphones. Why would that be different for the Corona App? 

If the goal of the Corona App is to protect the most vulnerable, then it is hard to see how an average somebody in his or her 70s or 80s will cope with this digital mean machine. 

15% - catch me if you can

Though there is another big stumbling block for the Corona App: its use is voluntarily.

We know that the “digitally savvy” Singapore saw voluntary download rates of 15% for their TraceTogether. In other words, one in seven users were convinced that a tracing app was a good and useful thing to have on their phones.

Because of its poor usability (the app cuts out when it is sent into the background) the actual usage rates will have been far below these 15%. Luckily, that situation is different in the German Corona App - or is it really? Jens Spahn’s major unique selling point in his latest press conference was that his app will of course allow you to listen to your music clips whilst the Corona App will happily beaver away without disturbing the user.

The grand total

Considering all these metrics it is easy to understand why one should be highly sceptical of the success of the app.

You simply can’t beat maths. 85% * 80% * 50% * 80% *50% * 15% = 2% (of the population).

Or put differently: if you are on a bus with 50 other potential virus spreaders, there is one person who has the Corona App up and running on his or her smartphone. 

Recovered already before you get the Corona warning?

As if this wasn’t enough, the Corona App is also flawed from a timeline perspective.

Now, again, as an app developer you would usually be asked to look into workflows and use cases, and how these relate to real life. I am not sure that ever happened when the German Corona App was conceived. 

On paper the workings of the app look great. It’s almost like a marketing brochure. When you are diagnosed with Corona, you update your status on your smartphone, everybody whom you have exchanged your secret data with via Bluetooth then is automatically informed and advised to go to the doctor.

But does it really work like that? The answer to that question is probably in most cases a resounding “no”.

Suppose you have successfully installed the app on your phone. At some point you feel sick and ask for an appointment for a Corona test. Your symptoms will show on average around Day 5. On Day 7 you can go for a test. On Day 9 you get your test results. On Day 10 you enter your test results into your app. Your contacts will be informed.

Likelihood is that those that you have infected already showed symptoms themselves, and are on the way to the doctor before the little Corona gremlin pops out from their phones.

#coronaapps #covid19 #sap #tsystems #robertkochinstitut