Management Information as a Control

What is it?

While process exception controls are great at spotting specific failures, most businesses have such complex environments that it would take thousands of these controls to provide enough coverage to be confident that everything is working as expected. It would also be necessary to imagine each of those thousands of things that could go wrong and detail controls for them. This would take a major effort and in most cases those possible failure points would never generate a material failure making such an endeavour non-commercial. You’d likely spend significantly more time and money documenting, building and then maintaining such a control environment than you would likely save from having those controls in place.

So a less targeted, more ‘zoomed out’ view, can be a highly cost-effective of providing broad coverage much more quickly and, with the right people available to review it, can provide coverage over items than would simply not even have been imagined with the process exception identification control.

The way to do this is to create a series of standard measures, tracked over time and importantly, in-sync with each other to provide context and highlight a range of possible failures.

How does it work?

By providing a series of measures, which all apply to the same time period or track the same cohort, it is possible to highlight changes in one measure which are out-of-kilter with the rest of the measures.

Here, we are not targeting a specific failure so we don’t need to apply the same level of imagination or detail over the specific failure we are looking for however a broad idea of the areas that present the greatest risk to the products or services that the organisation provides is useful. The broader the coverage of those measures, the more context is provided.

Management Information (MI) in this form can be easy to overlook as part of the control environment but it is important to stress that the MI on its own doesn’t achieve anything – for this to be effective, it must be reviewed and I would very much recommend having a formal review schedule or meeting in place where this is looked at and issues are raised and documented.

Example: Order Dispatch Pipeline

In this example, we will consider again an example where we are taking orders from customers and seeing them dispatched.

The expectations we set for our business are as follows:-

Customers can place an order online or by telephone

An order confirmation is sent out by email for all online orders.

Orders are put together at our one fulfilment centres.

Orders are dispatched within 5 days.

Orders are dispatched either by standard mail collection or via our courier where a one-day delivery service is paid for.

Orders dispatched by standard mail go via our franking department.

Orders dispatched via our courier have a printed label attached that we print ourselves.

We have a complaints system and around a dozen or so complaint reason codes.

By putting selected information side-by-side, we get an overall view of how things are running on a daily basis.

Looking at some of this data visually allows us to see problems occurring – take a look at the chart below showing orders taken, orders dispatched and the volume sat with the fulfilment centre.

If this chart were to show only orders taken and dispatched, it would not be easy to see an issue as they are broadly at the same level. The issue becomes more obvious however when we add the fulfilment centre numbers in and we can see that this is growing steadily over time – if something isn’t don’t to rectify the situation, this is likely to lead to a significant operational issue at some point in the future.

There are many ways MI can be used to provide this more ‘zoomed out’ view and this exact same data set and chart would allow a viewer to spot an issue such as order numbers declining or even dropping to zero on a given day. Even with this dataset that I have made up, you might reasonably question why order volume is zero on Sundays (even if our phone operation isn’t open on a Sunday, isn’t our website still up and taking orders? Do we have a problem with our website?).

It’s easy to dismiss this type of control as obvious but it is often poorly executed or overlooked entirely.

Advantages

· This type of control is great at providing broad coverage across a business or organisation and should be a good first step to establishing a control environment as it also supports organisational leaders in understanding what is happening inside their business.

· It is good at spotting broad issues and targeting effort to improve processes.

Limitations

· It requires some real mental effort to make it work. You’re pulling together a view of the world and you need to ask intelligent questions of the data and look at it in a number of ways. For it to be effective, it also requires a formal review meeting in place to ensure that it really is examined and that questions raised are followed up on.

· It is a lagging control – it tells you what has happened already and where things have gone wrong and this often isn’t evident until days or sometimes weeks after the issue.

· It isn’t good at spotting detail. If one or two orders have issues, it’s not going to be evident at all on a view like this. With a business delivering hundreds of orders, having a good customer service team to follow up when the customer calls through may be ok but in processes with more direct and severe consequences, this is unlikely to be enough on its own.

Some Examples

This control is good at spotting trends:

· Part of the operation is moving differently to another e.g. volume of cases growing despite constant input.

· More customers buying a particular product and complaints of a particular type (e.g. quality) have also risen but disproportionately.

· More products dispatched than we have had orders for.

This control is also good at spotting system issues:

· On a specific day, the number of new customers registered is zero when it is normally in the hundreds or thousands.

· Our billing system usually receives in the region of a thousand records per day from the order system but for the last week it has been around half that number.

The list here could go on as this kind of control is all about having a broad expectation from what has happened previously or from other parts of your business and finding things that don’t fit with that expectation.

First Published 30th December 2020

All views expressed in this article are solely those of the author

© Alex Cowhig 2020