Einstein Analytics - Using Data for Profitability Optimisation

May 1, 2020

Author: Axel Staal

In the age of data and technology, it can sometimes be difficult to establish a process which collates data from your existing (and potential) clients, along with management information, in a sensible and time-efficient manner while using it in such a way that it becomes vital to your business rather than becoming an onerous, time-consuming inputting task which causes more headaches than anything else. Salesforce, which already provides robust reporting functionalities on the data it contains, also has a very powerful analytics tool, namely Einstein Analytics, which allows you to delve further into your data and manipulate it to make it meaningful, with minimal input from your team, once the initial layout has been set out and implemented. So what does this look like, practically?

Very recently, we had the opportunity to work with Sionic to demonstrate the functionality and power of Einstein Analytics in transforming plain data (usually in the form of spreadsheets or database exports, from the likes of NavOne) into informative analytics dashboards that speak for themselves, by establishing easy-to-understand, interactive and accurate representations of real data (which was provided by one of their clients - a trust company). The core purpose was to enable Sionic to pinpoint where their client should focus their resources, in order to maximise their profitability. They can achieve this by easily viewing the performance statistics of their team members and the revenue versus cost of individual clients.

Click here to watch our full Salesforce Einstein Analytics demo with Sionic.

We were provided with two spreadsheets of anonymised data. The first contained information over 12 months in respect of the chargeability of each employee (grouped by team), which included the number of chargeable and non-chargeable units, along with various types of leave (contractual and otherwise). The second contained every invoice raised over the last 12 months, split by total work-in-progress (WIP) and actual bill, and included additional information such as fee types (fixed or standard), client types (key client or other) and to which team each client was allocated.

The files contained over 4,000 rows of data in total which, needless to say, is time consuming and potentially difficult to interpret and also has the potential to lend itself to human error... So what could we do to make this information talk on its own? Cue Einstein Analytics...

This chart below neatly summarises chargeability rates (minimum, maximum and average) by role and team. At a high level, we can see that Team 3 tends to have lower rates, irrespective of role. This may be due to ongoing projects or some form of restructuring, or perhaps Team 3 is just disconnected from the wider group and suffers from a lack in motivation. Whatever the reason might be, Team 3 could possibly benefit from some attention, and perhaps additional resources allocated to them in the form of a Trust Officer and/or a Trust Manager.

The heat map below goes into more granular detail, by showing chargeability rates for every employee in the business. Team 5, who generally have the lowest charge-out rates, are made up of bookkeepers and accountants; likelihood is that the preparation of accounts is included in the fee proposal at the outset of a relationship with clients, so, if that is the case, they can be forgiven and omitted from the analysis. The Trust Managers in Team 1, however, seem to be broadly in the red, as are the Senior Trust Officers; it may useful for management to investigate the reasons why. Similarly with Team 2, the Senior Trust Officers are broadly in the red and this could be something for management to look into.

In terms of leave, this scatter chart summarises absences by role which are then grouped by team. What is interesting in this instance is not the outliers in the top right hand corner (the bookkeepers and accountants in Team 5) but the one in the middle...

With Einstein Analytics, you can simply click on any point on the chart to dig deeper to try and understand out what is going on...

By clicking on an element of a chart, the whole dashboard changes in real time to show you what is relevant to your selection, helping you see the wood from the trees. This is a pretty neat, time-saving function which offers additional flexibility to the filters you would use at the top of the dashboard.

Turning to the theme of recoverability, the charts below show, firstly, the distribution of bills issued by each team which depicts a semi-annual bill run for the most part and, secondly, the difference between monthly WIP figures and actual bills which displays an overview of how much of the WIP is written off prior to the bills being sent out to clients.

Next, we can see the usual distribution pattern between key clients and the rest. What we learn from these charts is that there are a significant number of clients which provide very little revenue, and at times, none whatsoever, all the while using resources from the business. Whilst having a full book of clients is great, those which provide very little or no revenue could potentially be asked to look for a different service provider to relieve pressure on your team, and allow the business to shift some focus to higher yielding clients. Interestingly, Team 1 and Team 2 focus their efforts on key clients and have a large proportion of those on fixed fees - perhaps this explains part of the chargeability issue that was noted earlier on in this blog, namely the common theme that key clients are given significant resources and significant discounts.

Finally, we look at recoverability rates in what is probably the nicest visual representation of the data, in my humble opinion. What we can see from this is that there are a large number of clients which fall beneath the 100% recoverability rate over the course of the year, including some clients which demand huge amounts of resources from the business, while only actually being charged a fraction of the WIP. Looking at an example below, we can see that there is a client which had a £397k WIP, but was only billed around £140k... This is most certainly a key client, but perhaps this should prompt a conversation between management and the client to renegotiate the fee structure or extend the definition of out-of-scope work. Alternatively, should the client ask for a discount in the future, management can clearly demonstrate the huge discount they already receive, and justify declining such a request without causing the relationship to deteriorate.

There is a plethora of possibilities with Einstein Analytics and this was a brief example of what can be achieved. Some of the capabilities available in this software include artificial intelligence and machine learning to help analyse and explain the “what” and the “why” of past performance, as well as for forecasting future results based on the data that was inputted at the outset.

Bearing in mind that the data we were provided with was limited, I just wanted to give a quick overview of what this artificial intelligence and machine learning aspect translates into. The charts below, which were generated by Einstein Analytics Stories demonstrate how the software takes the data you upload, analyses it and gives you insight into what happened, effectively getting you from the start of the process (data input) to the end (analysis and results, as if you had gone through each of the previous charts and come to your own conclusions) with minimal input from users.

Of the 30 or so charts generated by Stories, I have selected three to illustrate the outcome...

This chart shows that Teams 1 and 2 outperform when the fee type is “Standard Fee”, as opposed to “Fixed Fee”.

This chart confirms what we discussed earlier, in that Team 1 and 2 outperform when the client is a Key Client.

In conjunction with the chart above, Team 3 outperforms when the client is not a Key Client.

For completeness, the datasets were uploaded in two ways: the chargeability section was input into Salesforce using custom Contact fields to contain the data while the recoverability data was uploaded by way of a CSV file.

For further information, or a live demo of what businesses can achieve with this tool, please contact us. We would be delighted to show you some of this in more detail!

Click here to watch our full Salesforce Einstein Analytics demo with Sionic.

More from Comnexa:

Watch a client success story: JTC Group PLC

Register for our next webinar: Operational efficiencies: how relationship data can help

Featured Posts