5 key tips for SME wanting to do more with their data

Recently I've been working with several small to medium sized businesses (SME) helping them to introduce a business intelligence (BI) tool to analyse and report on their data.

SME are different from larger organisations from a data perspective and so there needs to be a different approach to introducing a BI tool. Some of the key differences are;

• most of the time you're dealing with an owner of the business,
• most of the data is still in a raw form and still resides in the source,
• use of cloud software is increasing rapidly and not all of them have APIs,
• There's no real strategy or known end game from a BI perspective

With the cost of purchasing BI tools reducing so much it is much more realistic for SMEs to get into BI. Here's my top 5 tips for SMEs thinking of doing it.

1. Get a clear idea of why, what you want to achieve and what data you have

Understanding why you want to implement BI is important, you need to be able to focus your approach on the right outcome. Do you want to track and analyse stock movements? do you want to understand your customers better? do you want to track shipping times and delivery on time? There are so many things that any business can measure and report on that you can create a whole bunch of useless measures without a good plan and approach in place.

A couple of the questions I might ask my clients are 'how will that measure change the way you do things' or 'what action will you implement because of this measure'. These questions get the customer to think about the drivers of their business and what is truly an effect approach to KPI's and analysis.

Understand the data you have. This may sound obvious, however quite often the data within a small business is not set up to provide good KPIs or analysis. Any ERP or CRM that has been customised might be difficult to pull data from. For example; things like address fields that don't force conformity can make location analysis difficult and having customers in different systems with no customer ID can make it difficult to get an holistic view of a customer.

2. Work out how you will access your data.

Figuring out how you can access your data can help to speed up the process when someone like me comes in. Most software will allow for some way to access data. In it's most basic form some will allow you to export a csv or xlsx file of data or report. The more common now for cloud-based software is via an API. API stands for Application Programming Interface, basically it is a way for two applications to talk to each other and transfer data. These cloud software programs are usually termed SaaS, Software as a Service.

An API will require authentication, if you want to do some ground work you can search the help menu of the software to find out if they have an API then make contact with the support team to get details of any user names and passwords that will be required to authenticate.

Somewhere between csv/xlsx and APIs you may have a program that has a back-end database such as SQL or mySQL. Most BI tools will be able to connect directly to these as long as the tool has rights to, or is located on your network.

3. Do you want to automate, transform your data.

Moving to an automated process can be straightforward on simple data sets with good access but most likely this capability will start to move you toward the more advanced set up.

A simple set up might have one connection to a SQL database that is reloaded once a day.

A more complicated set up might pull from your main ERP system, a payroll program, Inventory control, GPS tracking plus a number of other possible data sources. When your business starts to look at this many data sets you will probably want to look at implementing a data warehouse to combine all your data into one place. This is often termed a 'Single Source of the Truth', which in essence means no matter what you use to view your data it should always give same the same output.

A data warehouse can be defined as 'a large store of data accumulated from a wide range of sources within a company and used to guide management decisions.'

This is a diagram we recently put together for a client on the process flow of their data.

4. Select the right tool

Selecting a BI tool can be a difficult decision as there are so many different tools out there on the market that all provide a slightly different set of capabilities and cost.

As a business owner or the person put to the task, you will probably look to bring in a consultant to advise you on this decision. When you do, keep in mind that most consultants will work with only one BI tool and will likely tell you it's the best one in the market. I've seen a lot of 'cult' like behaviors in the BI field around the choice of a tool. The reality is they are all good at something and not so good at other things. The important thing is; it does what you need of it.

To help with this decision there are a few organisations that research the available tools and make recommendations. The most commonly referred to is the Gartner Magic Quadrant.

Personally I can only comment on Power BI, Qlik and Tableau, as that's where I have experience, but I'm not going to recommend anything here. I truly believe each one is good depending on the requirements. The decision of which tool to use must be made based on your current and future needs and what suits you as an organisation.

Some key things to consider:

• How much data do you have
• Does it come from one source or many
• What is the access to those data sources like
• Do you need to create a data warehouse so that other system can access the data also
• How important is the visual aspect of the tool
• What cost can you afford
• Do you want a cloud service or a more advanced enterprise version.
• What level of security control do you need
• Do you want to automate the process and at what frequency.

5. Visualise what's important

Once you get to this point you will need to start thinking about what you want to visualise. At a basic level every dashboard / report / application will be made up of KPIs, Charts and Tables. These visuals will convey the data in a succinct and informative way so that you have the most key business information at your fingertips with the ability to analyse at a deeper level. Read my previous article 'Get MAD with your data’ for more information on how to structure a dashboard and report.

In summary

If you are starting the journey to introduce BI into your organisation, then I hope this article has helped. Be prepared for the long haul, it's likely going to take 12 - 18 months but you will be glad you did it. Take your time on it, focus on what you want to get out if it and use my favourite bit of advice.

Engage early, iterate often.