It’s hard to recall a more challenging time for businesses in New Zealand and globally. So when a business is asked to consider the importance of such organisational terms as data analytics, business intelligence, CAATs, data mining, continuous monitoring and a host of other names that sound great – but you have no idea what they really mean – then it’s no wonder businesses largely ignore them.
More importantly, with everything else on your company’s plate – rising costs, increased competition and overstretched resources – you feel that it’s not your highest priority at the moment. You’re too busy struggling with the day to day running of your business to divert your attention to something like data analytics.
But if any of those terms struck a chord with you then data analytics is actually the first thing you should be looking into to boost your organisation's productivity and performance. What you might not have realised is that answers to many of your most important business problems can be found within the mass of data generated by everyday operations.
Globally, the sheer volume of organisational data is growing immeasurably and any company that can better analyse, manipulate and make sense of this massive amount of data can use it to gain a genuine competitive advantage in their industry.
Why should you care?
At its core data analytics is all about turning your data into information in a form that makes it easier for you and your organisation to dissect and understand. The end game is that you can then facilitate better decision-making and deliver business value from facts. This is based on what your data is actually telling you rather than a gut feeling, or what your friend in the pub told you last night.
Some of the regular outcomes of data analytics are the identification of:
· Fraud, waste or abuse of resources that would have otherwise been unlikely to have been identified, or not identified for a considerable time
· Lost revenue due to errors and miscalculations in billing processes
· Invoices which have been paid more than once, allowing monies to be recouped from the vendors, and GST from Inland Revenue
· Inefficiencies in procurement processes, allowing savings to be gained
· Control weaknesses within business processes
· Errors in data conversion and migration processes when new accounting or business systems are implemented
As an example, a local company recently used data analytics to identify that several inefficient accounts payable processes were costing it tens of thousands a year. The organisation quickly implemented a few simple improvements and drastically cut down processing costs and improved its cash flow.
Setting your expectations
While it is nearly impossible to find earth-shattering insights from your data in just a few short hours, you can indeed confirm some amazing information about your organisation that you thought you were sure about, but were never able to prove 100 percent, and also quickly uncover many things you were never even aware of.
Performing a number of quick easy tests in a number of areas will give an organisation a good starting point for the analysis to quickly identify areas of interest or concern. Use a shotgun approach and ask: What are my spend patterns telling me? Am I doing things the most efficiently? Am I getting bang for buck with my customers?
How do you even begin to start answering these questions?
Step 1: Talk to your people - what are they saying and not saying?
The very first step is talking to your employees and getting to know what the issues and stumbling blocks are at the coalface of your organisation. Often you will find out about opportunities or obstacles you didn’t even know existed. Due to their granular and intimate knowledge of their areas, they will know the details best, and are a wealth of information to identify areas to look out for in the data.
Doing so will also give you a better feel for the data you will require with the added bonus of engaging with your employees and identifying where you are doing things well and where you may have a problem. This, in conjunction with your high level overview and priorities, allows you to direct the data analytics priorities, risks, or processes to be analysed first.
Step 2: Prioritise, prioritise, prioritise
With the initial legwork done you’ve identified risk areas, processes or controls issues and inefficiencies. This will ideally be used as the foundation of any analysis, which is going to confirm and quantify the findings of the areas being reviewed.
Brainstorming and engaging with your business analysts and front line managers will build up the tests / analysis you will finally want to perform. This extensive list of tests, or shopping list, should then be prioritised further, as there is often a risk of trying to run before you can walk. It is vital to take small steps and more importantly target the low-hanging fruit and get some nice easy wins under your belt first. It can be as simple as quickly and easily testing to identify vendors with incomplete contact details, conflicts of interests, or employees with high or negative leave balances.
This neatly flows onto the next step that will decide eventually what can or can’t be done. That is, can you obtain the information to support your testing? As the saying goes, garbage in means garbage out and you’re only going to be able to effectively analyse information that’s available to you electronically. For instance there’s no use attempting to analyse company fleet usage, if there’s no GPS electronic reporting available to do so.
This step will also quickly outline to the organisation gaps in the analysis and reporting process and identify potential future opportunities that can be prioritised.
Step 3: Tests away!
Having worked through with your IS and business units to identify the data you’ll need, you will then extract it from the various systems and pull it into the tool to be used to do your analysis. This is the main step where you are essentially converting your company’s data into useful, understandable, dissectible information. Technically, it involves data cleansing to ensure the data is standardised and to make it suitable for testing. Again, garbage in means garbage out. Testing means flagging your key criteria in the masses of data and then outputting the flagged records for subsequent examination and follow up.
Step 4: Were we on target?
Next, while it may appear to be a daunting process depending on the number of tests you performed, follow up on the results from your testing. Don’t give up hope; the effort is always worth it in the end. If you started small this shouldn’t be a problem, but if you went all in, you may very well be overwhelmed by the results and potential “false positives”. False positives are results which are incorrectly flagged as a hit due to underlying data layouts, rules or assumptions in your testing.
The key thing is that if there are a large volume of hits, that in itself may point to an underlying issue that you may have previously been unaware of. Once armed with this information, you can actually quantify issues and justify getting them fixed, or positively pat yourself on the back for not finding any skeletons in the closet. From GST variances to staff and vendors sharing bank accounts, the testing results will allow you to get to the bottom of things.
Your end goal here is identifying or quantifying issues so that something can be done to prevent it subsequently having a knock-on effect further down the chain. For example, duplicate vendor accounts increase the probability of duplicate payments occurring. These issues are frequently found at companies using data analysis and of course immediately impacts a company’s cash flow and GST returns.
Step 5: Optimise and try again
Now keep in mind it is rare that you will get a perfect outcome first time. However, acting on the initial results and dealing with the core issues means that when you re-run the testing further down the track and bench-mark to the previous results, you should see a marked improvement. Additionally, feeding “false positives” back into your testing should ensure that they do not come up again. This optimisation should mean less work the more frequently you perform the analysis.
It will also tangibly reaffirm that your organisation is proactively dealing with issues, rather than reactively, which tends to be a more time and cost intensive process in the long run. Using the earlier example of duplicate payments, in the long term, the effort and cost involved in reactively following up and retrieving such payments is significantly more than identifying how they are occurring and implementing a system control, or cleaning up duplicate vendor accounts to prevent it from happening again.
Data analytics is a great tool to add to your company's arsenal, however, keep in mind it isn’t a silver bullet and effort is going to be required to do it right. Like most things, you’ll only get out of it what you put into it.
If you are thinking that all this seems too hard and complicated, don’t. Depending on the extent of testing and amount of value you want to get out of it, data analytics can be implemented anywhere within several days using a shotgun approach, or several months for maximum effect. The number of opportunities where data analytics can be used is only limited by your requirements and imagination.
Data analytics is not a once off. Once implemented and ingrained in your company it’s much like a seedling. Your data analytics model will grow and mature as more data, processes, testing and analysis is added over time. The power of its automated approach will continue to provide you timely and effective reporting throughout your organisation for the life of the organisation.
Without prejudice, data analytics will allay any fears or concerns an organisation may have, and help answer all your organisations data questions and more.
Gladwin Mendez specialises in data management and data analytics at KPMG and has worked with over 200 organisations on a variety of data analytics engagements