Weeknote 30 March - 9 April 2020

On making trade-offs, when information systems do not mean IT, when numbers are not facts and when probability is meaningless. This time I have traded weeknotes for a fortnightnote^[1]. The rest of this period has also involved a lot of trades. Here are a variety of examples I have bumped into:

• Some teams and individuals have specialist skills which have become critical right now so they are having to postpone some bits or work or take on more help which then blocks other teams
• Many kids can’t go to school or stay with others so some of the team are having to work part-time or at odd hours
• Some work can’t be done remotely so we need the think hard about what risk we are asking people to take by doing it and if it is really worth it
• Sometimes working in a resilient way is also very efficient but sometime it means accepting a degree of duplication or over provision just in case
• People are often most productive when they can focus on what they are doing but we are having to interrupt some people alot as we uncover something important as part of our COVID-19 response
• Our suppliers are being impacted too which occasionally means having to wait longer for deliveries and, potentially, disappointing our customers.

Although the current crisis has made these examples explicit we’ve actually been making these sorts of choices all along but probably not thought about them as clearly. This might be something to hold on to when everything else gets back to “normal” along with less travel and more volunteering.

Most of my work over the past two weeks has been connected to the Council’s response to the crisis as almost everything else has been put on hold. It is probably not appropriate to go into too much detail on specific projects but some of the issues being thrown up will be worth thinking through later. Here are a few that have come up over the last few days.

Information technology vs information systems

Many services which are vital to people at these times are run by the Council. The current situation has impacted these in various ways such as changing the location (call centre staff are all working from home), volume (social care) or pace (website publishing) of services we already provide. There are also whole new things we are doing because of shifts in demand or central government policy. Our department is involved in all of this but sometimes in surprising ways.

You would expect us to be involved in getting phone and data connectivity in place for critical workers but we are specialists in information systems as well as information technology. A set of cross-functional “cells” have been created to coordinate operational work around a dozen priority areas. We already have the collaboration and remote working technology for these but we have also started to help streamline the flow of information and apply our agile techniques to help teams cope with the fast pace of change.

Numbers vs facts

Making decisions based upon data is generally a good idea but you need to make sure that the data is actually what it seems. We have all seen sensational versions of this (e.g. fake news such as the claims of links between 5G phone signals and the pandemic) but more subtle examples come up in our work which can lead us astray.

Take budgets and costs for instance. I have recently been discussing these for one of our largest programmes. These numbers follow each other around closely in meetings and reports. When the budget is bigger than the costs people tend to be fairly relaxed and there are lots of smiling faces. When they are the other way around tempers can rise. Unfortunately, it is usually a mistake to compare these and almost always a mistake to make a decision based upon a simple comparison.

A budget is an allocation of money and it is something that people can decide. Putting aside administrative errors, it is easy to determine what the budget is and rely upon it. Costs, on the other hand, emerge as work gets done and, although they are influenced by people, no one gets to decide what they are. Until you have finished the work you don’t know what the cost will be^[2]. Budget and costs are both presented as numbers but one is a verifiable fact and the other is, right up until the end, a subjective opinion.

Everyone struggles with this^[3] and a common response is to ignore the distinction and treat the cost estimates as if they were facts like the budget. Instead, we should recognise these for what they are and make risk-based decisions. A simple and effective way to help people do this is to develop ranges for the costs and make sure the ranges are presented whenever key decisions are being made. Unfortunately, a lot of management processes and behaviours actively discourage this. A precise estimate presented with confidence is often mistaken for competence when it should probably raise warning flags.

Probability vs Impact

When sending someone to work on site or making changes which can cut off someone’s network access can have such serious consequences almost all decisions become safety decisions. Risk management theory recommends weighing risks by multiplying the probability by the impact. I understand the theory, I really do, but it is almost useless in practice. To illustrate why consider these cases:

• We may need to increase the capacity of some of our systems to be ready for what the crisis might throw at us. Some of our systems are in the cloud so we can do this at the flick of a switch. Some are still in a data centre so I have to ask someone to leave their home to get this work done which, at the moment, could put their health at risk.

• One of our specialist teams need their network configuration changed so they can work effectively from home. The change is trivial to make, easy to check and only affects a few people but if the change uncovered a previously undetected bug it could bring down large parts of the network and impact thousands of people.

Probability times impact really doesn’t help us decide what to do. I’m just glad I don’t have to make the sorts of decisions our health and care professionals have to make every day.

A word about a good cause

There are a huge number of volunteer groups springing up to help people during the crisis. The engineer in me feels like these should be joined up into a coordinated system but, right now, hyper-local groups are probably beating the regional or national ones in terms of flexibility and pace. The COVID-19 Mutual Aid site^[4] has a pretty comprehensive list of groups and has post-code and map-based search facilities to help you find what is around you.

Footnotes: