Theunis Viljoen: There are some broad guidelines regarding financial reporting or issues that are found in most departments in organisations, whether it is underwriting, client acquisition, or investments.
Insurers have all this data locked up in various systems and the bigger the company, the more systems they have. An insurance company may have its investment in a very centralised unit, but a lot of companies still have autonomy of investments in subsidiaries across the globe. One of the difficulties that companies have is how to consolidate the results onto a common standard because a system may be different, and the people interpret the rules or the requirements in their favour. In other words, they will send you an answer to consolidate, but there will be a slant to it, and therefore you don't know whether the information that is coming from various sources has been prepared using the same optics and ruler.
"Many organisations have attempted to get rid of Excel and ban it from the
work desks of staff in favour of dashboards."
To counterbalance this, companies should create a common data transformation process, whereby the rules that are supposed to be used to interpret the data to transform it, are applied in the same way, irrespective of the country and without any manual intervention putting their spin on the data.
The second area that a lot of data and finance areas struggle with is Excel as the reporting tool of choice. Many organisations have attempted to get rid of Excel and ban it from the work desks of staff in favour of dashboards, but the reality is often that in their attempt to banish Excel, as a working tool or a modelling tool, they drive it underground.
People take data from their dashboards and copy and paste it into Excel spreadsheets. That leads to teams of people not doing important work and figuring out the tricky answers, but instead playing at decision making and reporting by spending their time on Excel spreadsheets. They feel that they have been busy when the answer is to automate these processes and spend their time analysing them.
They need to look at how much Excel is costing. If you took the number of people in the department and multiplied the number of hours spent clocking spreadsheets and factor in salary and overheads it can add up to huge amounts of money spent on an activity.
Theunis: We are talking about understanding the value chain of an organisation. It is based on the idea that you invest one pound in, that one pound will work for you through a process of parts of the business including investments. At the end of this process, your one pound has either become worth more or less, so understanding what exactly that value chain is, is the first task that the company has to figure out.
Then it is a case of making sure that you have the data optics on every aspect and then needing to understand what each activity in the value chain is contributing or detracting from your margin, and then figuring out which levers you have to pull to make that go faster or slower. These types of optics are about getting the data out of systems; being able to create that dimensionality.
Theunis: This links to the problem that companies have with using multiple systems.
What kills initiatives is trying to do point-to-point integration between the sources of data and the outcomes and the data that you are trying to achieve. If you have 10 major reports you want every month - a lot of companies create a Spaghetti Junction from all the data sources, and those could be different systems with different revenue streams. Then, all these systems have to be linked into one report. If you have 100 systems, and you have 10 key reports, you end up with 1000 integrations because each of those 10 reports may have to be integrated with 100 systems, and that is inefficient and costly.
You could integrate your systems into a data hub. The key thing is that by processing the data into this hub, there is a transformation that happens. That transformation is standardised for whatever type of data that you are looking for.
"Good management information is supposed to deliver unequivocal evidence
of what you should do - if it is ambiguous, it Is useless."
Doing it the other way around is a two-week exercise because emails have to go out to everyone to submit their spreadsheets and someone has to collate all the spreadsheets. You end up never sure about the quality of the information that you are trying to create and use to make your decisions to have maximum impact. It is like standing in front of the signal box as a Chief Investment Officer – you have measures that you can take to affect the outcome of the organisation, but you might not know what all centres pulling the levers are doing.
It should be critical thinking that goes into the process of decision making. Because otherwise, someone says help do something and you pull a lever and yes, you may avoid a crash, but it is causing a crash further on.
Good management information is supposed to deliver unequivocal evidence of what you should do - if it is ambiguous, it is useless.
Theunis: Some people like to look at data in tabular numbers form, and others like to look at it in a more graphic style. Data visualisation is still the final mile of the data journey. On the one hand, you've got this consumption of information, via spreadsheets and graphs. Then there are dashboards and the hope that creating a series of dashboards will help steer the company.
That is a fallacy because if that data is going through this manual process you end up with another manual process of trying to drive this dashboard. If you want dashboards with clear analysis then the way to reach that final mile of data consumption has to be to make sure that they are transformed in a way that if you ask a question, whether it is presented in a graphical format or a spreadsheet, it is still going to be the same answer.
"Critical systems thinking is always important - asking yourself if I knew
then what I now know, would I have made this step."
If you have different ways of feeding the graph and the spreadsheet, you are quickly going to end up with multiple versions of the truth. Insurance companies are not immune to projects embarked on even though the results aren't forthcoming. For instance, hiring consultants to come and build a great system, and maybe it doesn’t work, and following that project through because you don't want to accept defeat.
Critical systems thinking is always important - asking yourself if I knew then what I now know, would I have made this step. If the answer is no, then you should stop and start over again.
Theunis: Data is a two-edged sword. Proponents of data security say that you have to lock everything up. Security is important - we can't share it or show it to anyone; so, it is only if you are entitled to this piece of data that you can see it and the danger with that is then you end up with a lack of context. At a board level, you can only offer good advice, if you understand the full picture.
"Showing one person's achievements against another person's
achievement has benefits."
On the other hand, you have a school of thought that says share away, show people what is going on. Now clearly, there are aspects of data that are confidential – such as employee data, but we are talking about internal performance criteria.
Showing one person's achievements against another person's achievement or one department against another, allowing them to be able to create context about how well they are doing against their peers, has benefits. Let that data be useful in an organisation. There is more to be gained from a performance perspective if you share data widely in a company with due regard for making sure that it is not misused or leaked out of the organisation.
