Data Analytics Guide V2

Quick guide to data analytics

How to turn your data assets into customer insight to add value to your business

Quick guide to data analytics

Generate insight from your data

with 6 top tips plus a case study:

start thinking like a data scientist

You want to get your data to work harder for you and you want to use

the ‘data lake’ of customer information that you have stored; but you

don’t know where to start. These tips will help you to start asking the

right questions.

# 1: You already know more than you thinkYou probably already have a good idea of what you think is right and wrong with key areas of your business. You might even have something specific that you want to investigate. Just be prepared to learn as you go.

# 2: Get to know your data sourcesStart with a single data source that your business already knows really well. Check for obvious errors in your data – this is called Profiling your data. You can use free data profiling tools to do this if you have a lot of data. The other option is to export samples of your data to Excel spread sheets where you can explore the data sets in a familiar environment to help identify missing or nonsense data entries.

You can profile the data from each new source before you introduce it into your analytics reporting structure. As your understanding improves across each data source you can start to consider blending the data between the data sources.

“

”

Quick guide to data analytics

# 3: Keep it simple when you canIt isn’t always necessary to merge data sources and sometimes it just isn’t possible. If you already have reports from separate systems and you can compare the report outputs easily then you don’t need to integrate the data at source you could perhaps just produce an Excel spreadsheet.

Gaining an understanding of just what you can glean from the data available with the tools at hand is important and controls the scope of demands for reports from the wider business. Using systems that are already in place is the best way to start. The business may well learn so much from these first inroads into data analytics that it decides to invest further to gain more insight.

# 4: Think like a scientistWhen you fail you learn more than when you succeed. This might sound strange but it is important to be prepared to get things wrong. A scientist creates a hypothesis that is then tested through experimentation. Fact based evidence leads to a working theory that can then be used to create a conceptual framework.

As a data scientist your aims are to understand the relationships between the data in your organisation. You may start off with a hunch about a particular business issue; so consider what data sets surround the business issue process and then test your theories. Just remember to document the entire journey.

# 5: Fail fast and fail cheapAnalytics is a fast moving process and it is all about experimenting, documenting, learning and then moving on. Once the learning has taken place, the analyst can share the findings with the wider business, then move on to the next analytics project.

Quick guide to data analytics

# 6: Data specialists need to get out and about

Get the analytics team out to the different business departments and out to the customer so that they can be aware of data related issues and witness impact. Always remember that your data is your competitive advantage – it is a key asset of your business

Case Study: Online retail call centre based in South Wales

The call centre can easily track call volumes to establish the busy periods for their

Customer Service Agents (CSAs) so then they decide to develop their reporting by

measuring call volume by length of call and start to track if there are patterns

developing on length of calls at particular times of day. They use date and time, as

these data elements will be constant in each of their data source systems. The

telephone software they use also has a good reporting system that the business is

comfortable using.

Now the business decides it can match the stock inventory against the call centre

volumes to get an impression of the number of calls per sale, the number of items per

sale and the value of each sale.

So even though the two systems are not integrated they are able compare the data

from each source to plot productivity over different departments over one day. With

this information the business is then able to establish measures of activities against

each department.

Quick guide to data analytics

Immediate gain

1. The business insight that has been gained allows the business to plot trends

across its departments.

2. Once the business identifies these daily measures it can then make progress on

how to make improvements by assigning Key Performance Indicators (KPIs).

3. Putting in an analytical process that is making use of systems already in place is

almost always less expensive than creating new data warehouses.

Learning

The business realises through the data profiling exercise that there are frequent input

errors made by the call centre CSAs. Although the codified data input errors can be

resolved quite easily, it is not so straightforward with free text. If the customer service

CSAs have a data entry screen to input free text but then they forget to code the

complaint it will be difficult to analyse and learn from this vital customer interaction.

Data input errors can be rectified through better entry code design in consultation with

the CSAs.

What next?

> Download our Essential Guide to Better Data: over 25 pages of advice and

case studies on data management.

> Get weekly updates of all the latest data news from around the world when

you follow us on Twitter.