Data analytics is the analysis of raw data in an effort to extract useful insights which can lead to better decision making in your business. In a way, it's the process of joining the dots between different sets of apparently disparate data. Along with its cousin, Big Data, it's lately become very much of a buzzword, especially in the marketing world. While it promises great things, for the majority of small businesses it can often remain something mystical and misunderstood.
While big data is something which may not be relevant to most small businesses (due to their size and limited resources), there is no reason why the principles of good DA cannot be rolled out in a smaller company. Here are 5 ways your business can benefit from data analytics.
1 - Data analytics and customer behaviour
Small businesses may believe that the intimacy and personalisation that their small size enables them to bring to their customer relationships cannot be replicated by bigger business, and that this somehow provides a point of competitive differentiation. However what we are starting to see is those larger corporations are able to replicate some of those characteristics in their relationships with customers, by using data analytics techniques to artificially create a sense of intimacy and customisation.
Indeed, most of the focus of data analytics tends to be on customer behaviour. What patterns are your customers displaying and how can that knowledge help you sell more to them, or to more of them? Anyone who's had a go at advertising on Facebook will have seen an example of this process in action, as you get to target your advertising to a specific user segment, as defined by the data that Facebook has captured on them: geographic and demographic, areas of interest, online behaviours, etc.
For most retail businesses, point of sale data is going to be central to their data analytics exercises. A simple example might be identifying categories of shoppers (perhaps defined by frequency of shop and average spend per shop), and identifying other characteristics associated with those categories: age, day or time of shop, suburb, type of payment method, etc. This type of data can then generate better targeted marketing strategies which can better target the right shoppers with the right messages.
2 - Know where to draw the line
Just because you can better target your customers through data analytics, doesn't mean you always should. Sometimes ethical, practical or reputational concerns may cause you to reconsider acting on the information you've uncovered. For example US-based membership-only retailer Gilt Groupe took the data analytics process perhaps too far, by sending their members 'we've got your size' emails. The campaign ended up backfiring, as the company received complaints from customers for whom the thought that their body size was recorded in a database somewhere was an invasion of their privacy. Not only this, but many had since increased their size over the period of their membership, and didn't appreciate being reminded of it!
A better example of using the information well was where Gilt adjusted the frequency of emails to its members based on their age and engagement categories, in a tradeoff between seeking to increase sales from increased messaging and seeking to minimise unsubscribe rates.
3 - Customer complaints - a goldmine of actionable data
You've probably already heard the adage that customer complaints provide a goldmine of useful information. Data analytics provides a way of mining customer sentiment by methodically categorising and analysing the content and drivers of customer feedback, good or bad. The objective here is to shed light on the drivers of recurring problems encountered by your customers, and identify solutions to pre-empt them.
One of the challenges here though is that by definition, this is the kind of data that is not laid out as numbers in neat rows and columns. Rather it will tend to be a dog's breakfast of snippets of qualitative and sometimes anecdotal information, collected in a variety of formats by different people across the business - and so requires some attention before any analysis can be done with it.
4 - Rubbish in - rubbish out
Often most of the resources invested in data analytics end up focusing on cleaning up the data itself. You've probably heard of the maxim 'rubbish in rubbish out', which refers to the correlation of the quality of the raw data and the quality of the analytic insights that will come from it. In other words, the best systems and the best analysts will struggle to produce anything meaningful, if the material they are working with is has not been gathered in a methodical and consistent way. First things first: you need to get the data into shape, which means cleaning it up.
For example, a key data preparation exercise might involve taking a bunch of customer emails with praise or complaints and compiling them into a spreadsheet from which recurring themes or trends can be distilled. This need not be a time-consuming process, as it can be outsourced using crowd-sourcing websites such as Freelancer.com or Odesk.com (or if you're a larger company with a lot of on-going volume, it can be automated with an online feedback system). However, if the data is not transcribed in a consistent manner, maybe because different staff members have been involved, or field headings are unclear, what you may end up with is inaccurate complaint categories, date fields missing, etc. The quality of the insights that can be gleaned from this data will of course be impaired.
5 - Prioritise actionable insights
While it's important to remain flexible and open-minded when undertaking a data analytics project, it's also important to have some sort of strategy in place to guide you, and keep you focused on what you are trying to achieve. The reality is that there are a multitude of databases within any business, and while they may well contain the answers to all sorts of questions, the trick is to know which questions are worth asking.
While big data is something which may not be relevant to most small businesses (due to their size and limited resources), there is no reason why the principles of good DA cannot be rolled out in a smaller company. Here are 5 ways your business can benefit from data analytics.
1 - Data analytics and customer behaviour
Small businesses may believe that the intimacy and personalisation that their small size enables them to bring to their customer relationships cannot be replicated by bigger business, and that this somehow provides a point of competitive differentiation. However what we are starting to see is those larger corporations are able to replicate some of those characteristics in their relationships with customers, by using data analytics techniques to artificially create a sense of intimacy and customisation.
Indeed, most of the focus of data analytics tends to be on customer behaviour. What patterns are your customers displaying and how can that knowledge help you sell more to them, or to more of them? Anyone who's had a go at advertising on Facebook will have seen an example of this process in action, as you get to target your advertising to a specific user segment, as defined by the data that Facebook has captured on them: geographic and demographic, areas of interest, online behaviours, etc.
For most retail businesses, point of sale data is going to be central to their data analytics exercises. A simple example might be identifying categories of shoppers (perhaps defined by frequency of shop and average spend per shop), and identifying other characteristics associated with those categories: age, day or time of shop, suburb, type of payment method, etc. This type of data can then generate better targeted marketing strategies which can better target the right shoppers with the right messages.
2 - Know where to draw the line
Just because you can better target your customers through data analytics, doesn't mean you always should. Sometimes ethical, practical or reputational concerns may cause you to reconsider acting on the information you've uncovered. For example US-based membership-only retailer Gilt Groupe took the data analytics process perhaps too far, by sending their members 'we've got your size' emails. The campaign ended up backfiring, as the company received complaints from customers for whom the thought that their body size was recorded in a database somewhere was an invasion of their privacy. Not only this, but many had since increased their size over the period of their membership, and didn't appreciate being reminded of it!
A better example of using the information well was where Gilt adjusted the frequency of emails to its members based on their age and engagement categories, in a tradeoff between seeking to increase sales from increased messaging and seeking to minimise unsubscribe rates.
3 - Customer complaints - a goldmine of actionable data
You've probably already heard the adage that customer complaints provide a goldmine of useful information. Data analytics provides a way of mining customer sentiment by methodically categorising and analysing the content and drivers of customer feedback, good or bad. The objective here is to shed light on the drivers of recurring problems encountered by your customers, and identify solutions to pre-empt them.
One of the challenges here though is that by definition, this is the kind of data that is not laid out as numbers in neat rows and columns. Rather it will tend to be a dog's breakfast of snippets of qualitative and sometimes anecdotal information, collected in a variety of formats by different people across the business - and so requires some attention before any analysis can be done with it.
4 - Rubbish in - rubbish out
Often most of the resources invested in data analytics end up focusing on cleaning up the data itself. You've probably heard of the maxim 'rubbish in rubbish out', which refers to the correlation of the quality of the raw data and the quality of the analytic insights that will come from it. In other words, the best systems and the best analysts will struggle to produce anything meaningful, if the material they are working with is has not been gathered in a methodical and consistent way. First things first: you need to get the data into shape, which means cleaning it up.
For example, a key data preparation exercise might involve taking a bunch of customer emails with praise or complaints and compiling them into a spreadsheet from which recurring themes or trends can be distilled. This need not be a time-consuming process, as it can be outsourced using crowd-sourcing websites such as Freelancer.com or Odesk.com (or if you're a larger company with a lot of on-going volume, it can be automated with an online feedback system). However, if the data is not transcribed in a consistent manner, maybe because different staff members have been involved, or field headings are unclear, what you may end up with is inaccurate complaint categories, date fields missing, etc. The quality of the insights that can be gleaned from this data will of course be impaired.
5 - Prioritise actionable insights
While it's important to remain flexible and open-minded when undertaking a data analytics project, it's also important to have some sort of strategy in place to guide you, and keep you focused on what you are trying to achieve. The reality is that there are a multitude of databases within any business, and while they may well contain the answers to all sorts of questions, the trick is to know which questions are worth asking.
