How to Collect Customer Data
There are hundreds of ways to collect data, whether that’s through surveys, analytics tools or adding steps into the customer journey where customers have to share their data with you.
But, before we get into that, we first want to consider:
- Where in our journeys can we collect data?
- How can we organise and store that data?
- What are we doing to protect that data?
- How transparent are we being in collecting customer data?
- What benefits will we receive from gathering that data?
Of course there are other important things to consider, such as whether we are being compliant in collecting data, but these are really the key questions that we need to think about to put together a highly beneficial strategy for customer data analysis.
The first question of where to collect the data is particularly important, as Brent Bischoff a Cloud and Data Service Consultant at Business Systems says: “First, you need to look at all the places your customers are likely to go, and see if you can gain customer intelligence there.”
This is an important point to keep in mind as we present the following examples of data collection.
- Transactional Information – This involves logging a customer’s purchase history, subscription history and maybe even basket abandon data. We can collect this information through online tools or a contact centre advisor can log these things into an CRM system.
- Journey Tracking – We can track the customer journey and create data through many different means. For example, we can use click-to-call or click-to-chat links to see record where customers are bouncing into the contact centre. Another way would be to use tracking pixels, which are bits of code added to your website and emails, to track online visitors.
- Customer Forms – Usually online, these forms help us in the early stages of customer relationships to collect contact information. However, we don’t want to present long forms to customers at an early stage of our relationship with them, as this may harm business – but we can incentivise longer forms later in the relationship.
- Website Analytics – Smart phones and your website will be the route that most of your customers will likely want to interact with you. Fortunately, this makes data very easy to collect through tools like Google Analytics or Smartlook – so you can easily analyse your online demographic, as well as their behavioural traits.
- Contact Centre Analytics – By putting an analytics system in your contact centre, you can analyse customer conversations and therefore identify root cause problems, highlight training needs and even measure sentiment levels.
- Customer Surveys – Surveys are generally used to understand customer preferences and gather sentiment data. This sentiment data is then usually put into equations that allow us to calculate metrics like Customer Satisfaction (CSAT), Net Promoter Score (NPS) and maybe even Customer Emotion.
- Contact Centre Systems – Contact centre advisors summarise each call during wrap time, giving us lots of great data in the CRM system. But there are more contact centre tools that store lots of great data. Take self-service as an example, we can use data within our self-service system to find out where customers are abandoning the process.
Cleaning Customer Data
To be successful in our customer data analysis, the data needs to be accurate. So, we need to have a plan for data cleaning and validation, that may follow the process below:
No matter how we collect data, we will always run into outliers – we just need to know what and when to remove certain data points.
As Contact Centre Expert Dave Appleby tells us: “The key is to analyse the outliers that we remove and to also look for patterns within our data anomalies that might just be becoming business as usual.”
Another key consideration when cleaning data is not only what to remove, which is the more obvious part, but we also what to add back.
We can’t add in a zero into a complex set of data, as that seriously impact calculations that we make from the data. So, maybe we can add a mean figure or to maybe instead use a second or third deviation of the figure?
It will all depend on what you plan to do with the data and techniques that you use to analyse it…
Techniques for Customer Data Analysis
Now we have gathered and cleaned our data, we can start analysing and here are five customer data analysis techniques that will help you on your way.
- Regression Analysis – This involves examining the relationship between two data sets and looking for correlation, enabling us to assess how a change in one data set impacts another.
- Cluster Analysis – Finding clusters in data also allows us to segment our customers. For example, finding clusters in channel preference and customer satisfaction will allow us to group customers in terms of how we reach out to them and what we reach out to them about.
- Forecasting and Prediction – By analysing customer history, we can forecast future behaviour. So, if we spot a pattern in purchasing that usually leads to churn – for example – we can forecast that and make an effort to save the customer.
- Outlier Patterns – Not all outliers are one offs, they could highlight a slow shift in customer behaviour. So, by analysing your outlier you can find unexpected patterns and this will help you to find a root cause.
- Open Feedback Analysis – Not all data is numbers and we can look for keywords and themes in the open feedback that we collect, to find product improvement ideas. We can also look for narratives in amongst this feedback to understand how customers feel about certain things.
Things That We Should Be Analyzing
Now we have covered the essentials of customer data analysis, we can share some key examples of what customer data to analyse and how to make organisational improvements on the back of that.
1. Customer Demand
Most people don’t like to phone into the contact centre and unless talking to customers is a key competitive advantage for you, you should be seeking every opportunity not to talk to your customers.
This is where analysing customer demand comes in. Understanding why people are calling, when they’re contacting you and how you’re handling that volume can be a great place to start.
You will find some very interesting things by doing this and potentially save an awful lot of resource.
In fact, one contact centre – according to Duncan White – had to double in size from 350 FTE to 700 FTE in a nine month period because they didn’t understand why customers were calling – yet analysing customer demand would have given them a lot of the answers.
2. Customer Effort
The problem of analyzing customer effort in some contact centres is that you have to make your own mind up as to what constitutes customer effort in your organisation. Is it:
- Repeat calls?
- Wait times?
- On hold time / call transfers?
- First contact resolution?
The key thing here is that it needs to be something that can be recorded and measured accurately – with little effort.
Why is measuring customer effort like this so important? Well, it has a very close relationship with customer satisfaction, as highlighted in the Harvard Business Review’s well-known article: “Stop Trying to Delight Your Customers“.
This article revealed that when we lower effort, we can improve satisfaction and it also proposed an interesting way to do this, through analysing customer data and improving. They called it a “no audit”.
In the article, one organisation analysed how many times contact centre advisors were forced to say “no” to a customer, for any reason.
This led to them identifying many outdated processes and policies, which they were able to change, in order to massively reduce effort and it also saved them a million dollars over the first year.
3. Judgemental Heuristics
Judgemental heuristics is a framework or shortcut to understanding how people judge an experience.
One key part of judgemental heuristics is recognizing that one part of an experience is much more memorable than others.
As Duncan tells us: “Our analyses have always found that really high levels of customer experience are always correlated with a combination of task resolution, customer effort and judgemental heuristics.”
“So, it all boils down to; did you resolve the customers problem? How easy did you make it for the customer? As well as how did you make them feel and how did you leave them feeling.”
Therefore we should be analyzing touchpoints in our customer experience, to check that we are starting and ending a customer journey on a memorable note. We also want a memorable part in the middle, as in-line with the peak end-rule.
4. Customer Journeys
Customer journeys are the aggregation of all the customers activity that we need to understand, not just the point or the “silo” that you are responsible for.
There can be great value, in fact, of analysing how customers move from one touchpoint to another.
Let’s take a contact centre only example. If we look at the path of a contact through the contact centre and through departments can provide multiple insights into where bottlenecks are and where customers “fall-out” of a channel.
Another way we can analyse the customer journey and improve is to find out which channels your customers prefer to use for each type of interaction with your company.
“Understand why people use each channel. You can then provide the right resources to meet those needs,” adds Frank Sherlock, VP International at CallMiner.
“Also, you can point them in the right direction to use the right channel at the right time, resulting in happier customers, happier agents and reduced customer churn.”
5. Context Data
The nuts and bolts of customer data analysis is context data, which provides additional insight into an event’s, individual’s or product’s performance.
This is all about linking the context to the content. We don’t look at context data to find out what is going on, but to find out why.
For example, if we identify a trend where a certain contact type is always being transferred, we look at context data to determine:
- Is it because we’ve just released a new product?
- Is there a skill gap in the team?
- Is there an old process that is making it difficult to solve the problem?
Another example, in terms of the wider customer journey, is if we notice that sales in one location have stalled. We can then use context data to find out if it is because of competitor activity, less foot traffic in the area or something else.
We can then use context data to better understand the problem and make more informed decisions to increase revenues and improve customer experience.
Customer data analysis involves collecting, cleaning and grouping customer data, before looking for trends and key areas in which to improve organisational performance.
Yet, there are many different types of customer data to collect, including personal data, engagement data and sentiment data.
The key to deciding what to analyse is to have a purpose before you start your analysis. This purpose must be to improve in one area.
Just remember, if we are not going to find the time to improve, what’s the point in doing the analysis?