What is Transactional Data Modelling and what can it do for your business?

The modern business is awash with data. Every digitised system you run, whether it’s a website, an electronic point of sale (EPOS) platform, enterprise resource planning (ERP) or customer relationship management (CRM) software, it all generates constant streams of data relating to various aspects of business performance. 

It’s no longer any great secret that the ability to harness data and turn it into actionable intelligence is a cornerstone of commercial success. The global analytics and Big Data market is worth almost £150bn.

It costs money to transform volumes of unstructured raw data into meaningful business intelligence. It requires sophisticated processes of modelling, interpretation and analysis using advanced algorithms. But business leaders are ready and willing to spend big on data analytics because, when you base strategy and operational decision-making on clear insight about what is actually going on in your business, the returns are significant.  

One report found that data analytics was generating £12.20 for every pound spent – an enormous ROI of 1120%! 
A typical business generates all sorts of different types of data, and the various categories of data have various impacts on performance. A lot of attention gets placed on customer data. After all, the better you understand your customers, the better you can target your offer to match what they really want, thus boosting retention, repeat business, average spend etc. That’s one of the first things you learn about marketing.

But while information about who your customers are is important, what really drives value for a business is understanding how they behave. And the best place to find this is to look at purchase transactions – what is being bought, who is buying it, when things are bought etc.

Ultra-detailed insight into customer behaviour is just one reason why businesses are turning to transactional data modelling, though. Data from sales processing can be put to a number of other uses. And customer purchases aren’t the only type of transaction that takes place in a business.

What is transactional data?

Transactions take place right across your business. Every purchase a customer makes, every order from a supplier, every time money is transferred into or out of one of your accounts are obvious examples of transactions.

But beyond purely financial, you can also count every order fulfilment (i.e. when goods actually exchange hands), every time a task moves from person to person or point to point in a workflow, or even every digital interaction – every click, every web page hit, every comment, like or product added to a cart – as a type of transaction.

As long as they take place on a digital system, each type of transaction will be recorded as a unique set of data. So data is readily available for financial transactions, operations, logistics, digital and more.

One of the defining characteristics of transactional data is that it is very specific. Take a purchase a customer makes, for example. Your online shopping cart or in-store POS will record the full transaction amount, a breakdown of the items bought, the time and location of the transaction. If the customer pays through an online account, loyalty scheme or even with a card or digital wallet, your systems will also capture details about them.

Every single transaction therefore gets logged in a great amount of detail – the time, date, type, who was involved, what goods were involved (if relevant), the value, any special conditions (discounts, redeemed vouchers etc), and so on and so on. That’s a huge resource of information covering what happens right at the coalface of your business.

By delving into this data, you can start to identify patterns hidden deep in the fabric of what happens across your business that would otherwise remain invisible – which groups of customers respond best to which marketing messages, what products they prefer, when they like to shop or browse most, how peaks and troughs in sales correlate to time of day, month or year, the impact of pricing and promotions, which products trend when and with whom.

All of this provides an invaluable resource for optimising sales and marketing strategies, driving operational efficiencies, creating customer journeys/sales funnels that maximise conversions, and ultimately improving your bottom line.

What transactional data can be used for

Let’s take a closer look at some of the things transactional data modelling can do for your business.


Because of the level of detail transactional data contains, it lends itself perfectly to predictive modelling – using patterns identified in historical data to make predictions about future events. Predictive modelling, also known as predictive analytics, is all about calculating probabilities from known events. The more known events you know about (i.e. the more historical data you have access to), the more accurate the probability calculations will be.

So if you take a year’s worth of sales data, for example, and run it through a predictive data model, you can use what has already happened to make forecasts about overall revenues, peaks and troughs in trading over the course of a calendar year, staffing requirements at different times, the impact of different marketing campaigns or promotions on footfall and sales etc. If you have five or 10 years’ of data to draw on, or a large resource of industry data drawn from similar businesses to yours, the predictions become more and more accurate, allowing you to optimise operations in advance with increasing confidence.


As already mentioned, arguably the most common way data is used in the modern business is to gain an in-depth understanding of customers – their characteristics, their habits, their expectations – in order to create the strongest appeal to the most people possible. Because of the level of detail it offers about actual customer interactions with your brand, transactional data is a powerful tool for improving your relationships with customers.

In particular, it can be used for:

  • Personalisation: Knowing your customers’ purchase history and habits is especially powerful when it comes to personalising marketing and promotional campaigns. That can mean differentiating messages and offers depending at which stage of the customer journey they are at (see below), sending communications at different times based on when your data suggests different people are most likely to take an interest, or simply building offers around past purchasing preferences.
  • Propensity Modelling: Knowing how likely a customer is to make a purchase depending on whether you target them in X, Y or Z ways is not only useful for personalisation. Ultimately, some customers are much less likely to buy than others whatever you do – are they worth the time and effort, or could you focus your energies on better leads? This is what propensity modelling does – based again on known customer behaviour, it tells you the likelihood of customer A or B converting following a range of possible interactions.
  • Customer Journey Mapping and Retention: People’s relationships with brands change over time. They might start out as one-time, occasional or reluctant buyers, but over time become more regular and loyal customers. Customer journey mapping uses comparisons of individual’s shopping histories to identify the relationships between low value, disengaged customers and your most loyal followers, what things trigger lower value customers to start purchasing more, what common factors influence first-time buyers to become long-term customers etc. By looking at customer relationships as a continuum where certain inputs influence them moving to the next stage, you can use your customer journey ‘map’ to target the right activities at the right people at the right time, ultimately driving up retention and customer value.



Finally, one thing that businesses (or at least marketing departments) have long struggled with is measuring the impact of campaigns in a way that is reliable. Prior to the digital era, all sorts of weird and wonderful metrics were conjured up to justify marketing spend.

But thanks to transactional data, nowadays we have something that is concrete, factual and accurate. If a customer clicks through to buy an item from a Google or Facebook ad or an email shot, if they sign up for a loyalty scheme or present a discount QR code in store, all of this is logged as part of the transaction. You can measure directly the impact campaigns have on sales, allowing marketers to quickly identify what really works and what does not.

Transactional data modelling is just one of the areas of data intelligence that we specialise in at Key Element. If you want to find out more about how we can help deliver greater value from your data, please contact our team.
Scroll to Top