8/10/2020 by Theo Goutier

What are we (still) blind for with Covid-19 data (and in our own businesses)?

Following our series of blogs on analysis of available data on Covid-19 to show what data can, and also cannot, tell us we will conclude more in this blog.


Please note: Our intention here is to show what can be done with a comprehensive data set, frequently available but under-utilized in companies, not to replace experts in providing advice on Covid-19. The reality of Covid-19 is far more complex as health-related, behavioural, political, social, economical and environmental factors all play a role and impacts a lot of people negatively. Our sympathy is most with those affected with their health or loss of their loved ones. All we can hope for is that some of what we found to be hidden helps those protecting each of us and what we can do ourselves.

In our last blog early September https://www.dobilo.com/news/blog/2020-sep-2nd-data-blindness we said we have been blind with the Covid-19 data available to us.

To be honest we said why but not how blind we have been. In the end, as with any extensive data-analysis, we need to dare drawing conclusions and to take the consequences too. To illustrate what most likely has happened look at this short video:

Based on the worldwide death-rate with Covid-19 patients, we can derive that in Netherlands and Belgium, and many other countries, the number of people that did catch the virus were 5-20 times higher than those found in (limited) testing back in February, March and April.


Our main source of data used on Covid-19 explains this in a balanced way https://ourworldindata.org/covid-mortality-risk "What we want to know isn’t the case fatality rate (CFR): it’s the infection fatality rate (IFR)". Today the CFR is 1.05/38.5 million worldwide or 2.7%. IFR is definitely lower and scientists estimates range from 0.6% to 1.4%. 


This factor of '5-20' times more unfound cases vs reported seems to be substantiated by the official Dutch Government Covid-19 Dashboard and its Health Agency RIVM with alternative data and even in 3 ways:

 1) General Practitioners / common doctors report people who have Covid-19 type of complaints 



2) Measurement of virus-RNA found in the sewer systems In various Dutch cities


3) The RIVM itself estimates how many were contagious on a particular day and the estimates back in March, at the peak of Wave 1, are up to 2 times higher than today (Oct 7th). Official reports stated about 1300 found daily around the peak. People are contagious for about 7-10 days so 10-13.000 were found of 200.000 minimum out there which is 16-20x more.



As in one of our previous 'Covid-19-data' blogs on how to perform proper data-analysis, gauging the data found one way with another way, above proves this data 'blindness' when just counting positive tests.

We can thus conclude with caution, across the whole set of blogs, and other sources used, the following: 

· The virus was widely spread in Western Europe in February already around 20x more than reported positive. (very likely) 

· The SARS-CoV2 virus is part of one family of the 200+ common cold viruses. 

source: https://en.wikipedia.org/wiki/Common_cold

The common cold is a viral infection of the upper respiratory tract. The most commonly implicated virus is a rhinovirus (30–80%), a type of picornavirus with 99 known serotypes.[29][30] Other commonly implicated viruses include human coronaviruses (≈ 15%) .....

· This new cold-virus effects some in a more aggressive way. The weakened, highly aged and some with certain (genetic?) susceptibility in which the immune-system over-reacts (respectively very likely to likely) 

Those getting hospitalized or died were just the tip of the first wave (iceberg). Also the others ill or just positive without symptoms were part of that tip. Millions more people in Europe have been 'infected', most without even knowing it or knowing it for sure as they were not tested.


Conclusions we can more or less likely draw too:

  • The current wave is still lower than the first as we increased testing but 4-5x more people are still catching the virus daily (quite likely) 

  • In this (relatively hot) summer in (Northern) Europe, the virus was less active not to say dormant (likely)

  • Lockdown measures helped in spring to flatten the 1st wave (likely)

  • Keeping a bit of distance outside helps although much less people are catching it outside (very likely)

  • Temperature/moisture conditions (climate cycles regionally) seem to be the main driving force behind the waves (quite likely)

  • Most will not get very sick from this virus, like with other cold-viruses. Majority of people most likely are resistant or will not suffer or if they do not for long and severe. (quite likely)

  • Contact-tracing is not so effective if the virus is out there with 4-5 persons that are not found versus 1 found with testing positively. (likely)

  • The vaccine will help the weakened persons most. (likely).

  • Relatively reliable (<20% false outcomes) speed-tests when cheap and made available can become the same as brushing your teeth: 1 test a day keeps Covid-19 away (from others).

Now ‘finally’’ we can talk about the consequences too:

  • Rapid development and deployment of vaccines are needed. First to vaccinate the weak(er) and elder ones like we do with the flu vaccine yearly

  • Speed-test production and roll-out to all schools, nursing homes, hospitals, sports clubs, households etc. should become the focus now

  • Meanwhile, ventilate and keep distance inhouse mainly

  • Wear masks, whether non-medical ones help or not but at least it keeps us reminded of the virus still being out there and to keep that distancing alive 

  • Wave 2 could very well become much bigger but the deadly effect is already lower. We have fortunately been able to learn how to cope, now it is time to learn more and more quickly . Let us protect the weak even better, as that's what we have learned most with Wave 1

For us running businesses: what should we take out of this Covid-19 'data'-case series of blogs?

Do careful analysis and frequently circle back to what truly is happening as this is not always directly visible from the data at hand. Use these steps: Observe, Observe and Observe some more, Conclude with caution, take Actions, starting with experiments, to Observe again what is the impact. Keep in mind that there might be factors out there you have not witnessed i.e. have been blind for so far. Don’t let emotions make you conclude too soon or based on biased observations and take wild actions that are most likely counter-productive.


To illustrate this further we can highlight the case of a warehouse handling full pallets for distribution in the UK from Belgium. The company had implemented a new Warehouse Management System (WMS) with barcode scanning with the main goal to improve the labour productivity. However our help was called in as productivity had plummeted after 4 months with 25% instead.

Although there was detailed data for every pallet-movement since Go-Live, there was no clarity on what had been driving the decline. Nor did this vast data show clearly where in the process of handling in, storing, picking and handling out the decline occurred. Simply, because data was not captured as performed as the forklift-operators all had adopted different ways of dealing with the hick-ups and difficulties they had experienced with the new system.

Hence we did a daily count over two weeks, via ticking boxes on paper A4's on the forklifts, of various issues we had heard from the operators. Out came the following:

  1. time to process a pallet, via the hand-scanner, frequently took  longer than 3 seconds
  2. the system asked to pick a unique pallet that was stacked at the bottom of a pile of pallets of the same SKU while before operators took the easiest pallet from the stack which thus was taking a lot of extra time
  3. pallets had to be put away after handling in, at locations far from the unloading dock.

For each of these top 3 causes a further investigation of the deeper lying causes and solution was put in place. Productivity came back to the original level after 3 months. However it was also decided to find a simpler WMS, on the longer term, as the connection time for the scanners with the server was still too long.

Let us at Dobilo also help you on observing the truth of what your data tells you, or doesn't in many cases, and help run your business based on the right observations and conclusions faster, better and with more control over what you could and should control. Contact us via below form.