Adopting Analytics for the SME

Decision support systems were, until recently, only available to large enterprises. The complexities and costs associated with business intelligence or analytics solutions made them prohibitive for other companies.

Smaller businesses had to rely on standard reporting built into their operational applications or, at best, employees had to export data and indulge in complex data mash-ups within the limitations of solutions like Excel. Reporting resources and time, be it directly from end users or vendors building the capabilities into the applications, focused on provisioning mandatory reporting for financial or industry regulators. A business would largely survive on the skill and instinct of the management due to lack of data-driven insights, limiting growth to the bandwidth and skills of that leadership.

Improvements in business applications, databases and technologies such as the cloud, memory database performance, and IoT offer the means to build out the value of reporting and decision support. To realise such value, business owners and their advisors have to assess how and what data should be captured and to what standards and quality.

A professional services company needs to understand their performance by the service areas provided, their type of clients and projects. Retailers, for example, will want to understand performance against product categories, brands and stores. The initial step in any analytics journey is to define these dimensions and measures, and then more importantly design how to capture them accurately and consistently within the processes and transactions of your business. This is an often-forgotten challenge in Business Intelligence and Analytics; the solution cannot be built in isolation from the core business and processes. Captured “measures” such as revenue, direct costs and overheads, must be linked with business dimensions in a structured way. For instance, a retailer recording sales by outlet would store this dimension with a structured “identifier”, so the data can be specific and historically consistent.

While legacy ERP, General Ledger, or similar transactional systems allow capturing of quality structured data in a consistent manner, historical transactional systems often pose a further challenge to restructure and transpose that data due to the finite resources available to “on-premise” systems. Data is traditionally optimised to support transactional processing, allowing it to be added, deleted, or updated at a granular level, but it fails to support timely aggregations or other on-the-fly calculations. Hence, traditional BI solutions lay separate to the business execution transactional systems, and often represent a delayed picture of the business.

This challenge is, to some extent, addressed by vendors who transition their platforms to the cloud. Numerous clients running their business on common cloud infrastructure and solution vendors that seek to differentiate themselves, have seen cloud-based business solutions blend analytics into their functionality to solve common problems. This is enabled by virtue of a more powerful infrastructure at the disposal of the cloud vendors, relative to what had previously been available to the client and their “on-premise” solutions.

Click to enlarge

In such ways, decision support becomes ingrained within the operation. This is further supported by modern browser user interfaces, exploiting HTML5 type technologies and thereby embedding data, information and KPIs into the very tiles and buttons from where transaction execution of functionality and / or a deeper review of the data is provided.

This aggregation of transactional data and analytical data into one source also brings another tangible benefit as analysis goes “real-time”. A retailer’s POS, Inventory and Finances stay aligned, provisioning up to the minute assessments on how a new store layout is performing, and against previous benchmarks. Success can be rolled out to other stores within the day, failures rolled back.

Add or “mash” into your financial data, external data sources such as loyalty programs and in store footfall counters and you can assess the success of your different channels and their cross-adoption, or what your conversion rate is from people entering a store to those making a purchase (a basket). The cloud also allows you to bring into the analytics domain 3rd party data, whether assessing seasonality / implications on sales or tracing social media sentiment across your brands.

Like retail, every industry has its own KPIs, data and key questions to assess. The process, however, remains the same.

You first need to assess how you want to understand your business, what data do you need to capture and retain to conduct this assessment. How do you maintain quality and consistency of that data; where will you put that data and how to structure it to make it accessible? How can you deliver such data to the right people so that they can investigate, gain value and make the right decisions for your business? All this may have implications across your whole IT landscape.

How do you establish the correct architecture, not only to support execution within the business as it is now, but to help anticipate changes and build the business you want tomorrow?

This article was originally featured in The Accountant publication and has been reproduced by kind permission.

Did you find this useful?