Chapter 14
Analyzing Data to Drive Results
- Realizing that data can be extracted from social media
- Understanding that data goes beyond the numbers
- Leveraging data to build deeper customer engagement
- Planning for the ever-evolving social data
There's no shortage of information and data created through social media. Nearly every move we make online is traceable, and trackable. Every time a user likes, shares, or comments on a piece of content on Facebook, it generates a data point for Facebook itself, and for the content originator. Twitter users know when their tweets are retweeted, and Twitter provides analytics for reach and penetration.
For businesses, this data is a jackpot of insight into the social customer. This social data illustrates a customer's preferences, behaviors, views, ideals, and so much more. CRM 1.0 involved collecting basic customer information: address, phone numbers, birthday, place in a sales cycle, and so on, but social CRM data gives new life to your understanding of your customers. You get to know them on a more attitudinal and emotional level. You learn their sentiments.
In this chapter, we show you the best ways to harness the massive amounts of customer data generated in social media, and we help you identify ways to filter the good information from the less relevant data points. Finally, we show you ways to leverage gathered data to enhance customer experiences with your brand.
Understanding the Social CRM Data Storm
Information flows in many directions in social business. It isn't just companies shouting through a bullhorn to customers. Consumers engage brands on customer-preferred platforms, and then other customers, vendors, partners, and employees join the conversations. This provides a very different view of customers, one that is closer to a 360-degree view.
This is all great news for business, albeit at times a bit overwhelming. A tremendous amount of data is gathered every day, and not all of it is useful.
In order to sort through all the social data available and turn it into useable metrics, businesses need social CRM monitoring and reporting tools. These solutions allow businesses to harness their data and translate it into a greater understanding of the social customer.
Filters and processes can help you sort what's relevant from what's not. To do that, take a moment and think about what you would want to know about your customers, in a perfect world with no data restrictions. Identify the key traits or features that you need to get a clear and accurate picture of your social customers. This varies by industry, and may change within your own business. For instance, consider the following examples:
- When you're rolling out a loyalty program, you may want to know who your most engaged customers are. You could look at how many sales can be attributed to current customers, or how many times customers share your messages in social media.
- During back-to-school season, you might then want to know how many customers have children, and their ages.
- When you open a new bricks-and-mortar location, you'll need to know which customers live in that area.
All this information is available through social channels. You just have to know where to look, and how to get it.
Plenty of tools can help businesses mine data from the social chatter. Your options include the following:
- Each social network has its own innate reporting capabilities, such as Facebook Insights.
- Social CRM solutions, which we cover in Chapter 16, also include reporting.
- Sites like MediaVantage (www.mediavantage.com) and Track Social (www.tracksocial.com) have taken data collection and analysis to a new level. These sites can pull all the data into one place and enable businesses to customize analytics and reporting formats for their needs.
No matter what tools you use for reporting and data analysis, make sure that you're looking at data within the context of your business structure. Determine relevant metrics to help you reserve resources and focus on the social conversations that hold the greatest value.
Teaching the Different Parts of Data
When determining what data you want to corral, it's important to keep in mind the different types of data and what they can teach you. Many businesses use historical data to take an educated guess at future performance across similar metrics. Forecasting, budget planning, and sales goals are often based on historical analytics data, but this approach can be limiting. To capture a well-rounded historical view for planning, it's important to understand the difference between descriptive analytics and predictive analytics.
- Descriptive analytics: This type of data identifies past events and the perceived or real factors that played into creating the events. The idea is to look for indicators of success and failure in past initiatives. It's a reactive strategy to marketing planning — looking back, reacting to discoveries, and adjusting future plans to meet the past conditions.
- Predictive analytics: This is a mathematical model that accurately predicts future results and outcomes based on hard facts. It's easier (and typically more affordable) to keep existing customers than to generate new ones, which is why predictive modeling often leads marketing initiatives for promotions and offers. Taking what you know customers prefer and reaching them with what they want falls into predictive modeling.
Keep in mind that the past can't always predict the future. Both descriptive and predictive modeling have their limits, in part because the following factors make tomorrow (or next year) a little different from yesterday (or yesteryear):
- Today's marketplace is ever-evolving and complex.
- You're gathering a lot of data at a rapid pace in social CRM.
- Markets are moving targets that you can't always accurately predict.
However, if you analyze historical data in the right context, you can develop a more tangible approach to planning. But it's vital to distinguish the types of data you're gathering and analyzing first, as we outline in Table 14-1. After you know what you have, you can draw conclusions and translate that data into actionable metrics.
Table 14-1 Key Differences between Descriptive and Predictive Analytics
| Descriptive Analytics Modeling | Predictive Analytics Modeling |
| Perceived outcome based on historical data | Accurate prediction of outcomes based on past outcomes with the same factors in place |
| Relational | Equation-based |
| Rearview | Future-focused |
| Reactive | Precision planning |
Combining Business Intelligence with Social CRM
Taking business data and turning into actionable knowledge is business intelligence (BI). It's seeing your organization for what it really is and what it has the power to become by considering all the data available. BI technologies aim to empower an organization with better decision-making processes, using the power of knowledge at its fullest to create a competitive edge for your brand. In many ways, it's the precursor to the kind of data we have access to through social CRM.
BI technologies provide historical, current, and predictive data for any number of business purposes. Here are a few ways that you can use BI to direct and affect brand initiatives, along with social CRM:
- Reporting: Say your marketing team is leading a campaign in order to launch a new product. Your social CRM or BI system can pull data about online conversations, such as retweets, while your BI system tracks sales. In this way, you can see the success of the campaign through social CRM and BI together.
- Forecasting: The same marketing campaign mentioned in the preceding paragraph can also be used to predict the success of another new product launch later in the year. Descriptive analysis will tell you what types of messages were shared more often, and social conversations can be tailored accordingly for optimum reach.
- Strategic planning: To determine what your business's next new product should be, look to predictive analytics to determine what consumers want. By looking at the questions customers asked during the most recent launch, you can get a clear picture of what features customers were hoping for the first time around.
- Performance measurement: Make BI and social CRM reports easy to consume, with clearly outlined goals and metrics. To continue with our product launch example, you might set a goal of 100 retweets (RTs) and 1,000 units sold on day one. You can then easily track this and report on performance through BI and social CRM.
Structuring Data Collection and Reporting
Before you begin gathering customer data, be sure you're moving with purpose. Avoid wasting time and resources by targeting your data collection and reporting efforts. Ask these questions to guide what reporting you truly need:
- What targets do you want to track for your business? Consider whether you'd benefit most from looking at all your customers, or just a core group.
- How much flexibility and customization do you need in your reporting? Decide whether you'll be interested in the same metrics every day, month, or year, or if it varies seasonally.
- Will you use reporting to gauge employee performance? Evaluate whether you can or will tie social metrics to specific employees. For example, if coworker Matt is mentioned on Twitter, how will you use that information?
- How often do you need to update reports? Think about what types of data you can realistically compare. Daily data is a lot to digest, but yearly is too long to wait. Many businesses find it helpful to look at summaries weekly and full reports monthly or quarterly.
- What is the best way to disseminate the data and intelligence? Identify who needs to know what, and when, as well as the best way to present information. Some information may be helpful for everyone, while you should reserve other information for management.
- Who needs to be abreast of data reporting? Appoint a few people to analyze the information and disseminate the reports they create. Not everyone needs to see the raw data.
- Do you need to set threshold alarms if certain changes occur with data points? Imagine the best- or worst-case scenario. What would you want to know the moment it happened, without waiting for a weekly or monthly report?
- How quickly do you want reports generated? Create a timeline that works with the cadence of your reports. For example, a weekly report will need to be available more quickly than a quarterly report.
- What amount of time is reasonable for employees to spend generating reports? Allocate time efficiently. Reporting is important to track your company's health, but not at the expense of work that leads to trackable data.
Make sure you're collecting and storing only the data that's most relevant to your business goals. Determine what you need most and shape processes to fit that efficiently.
Translating Social Media Data Into Metrics
When many of us first think of data and analytics, we think of quantitative results suited for spreadsheets, and we can yield powerful insights by analyzing quantitative factors. The number of RTs on Twitter can tell you how many people found a piece of content interesting, and the number of likes on a Facebook Page provides a good indicator of people who want to engage or associate with your company.
However, numbers aren't the whole picture of circumstances or economic environments. What about valuable words within text? Next we tell you about the value of text analytics and text mining.
Defining text analysis
The process of deriving high-quality information from text is often referred to as text analysis or text data mining. Patterns and trends are applied to text to identify a value. Spam filters exemplify a widely recognized use of pattern recognition and text mining. Your e-mail service provider taps into linguistic features that typically indicate an unwanted bulk message.
In social CRM, sentiment analysis looks at the text of a customer's message to determine if it's positive, negative, or neutral. Like spam filters, sentiment analysis isn't 100percent accurate, but it can provide a strong indicator of trends. A sharp spike in negative messages is worth looking into, even if the exact number of messages is off slightly.
Text mining is a complicated process that goes on behind the scenes of most social CRM solutions. It starts with a classification system using statistical, linguistic, and algorithmic techniques to analyze text. The information gathered from that text is then arranged in a format that's easy to read and provide metrics on.
Turning textual data into high-quality information combines a set of guides for relevance to the entity conducting the text mining. Here are the typical ways text is consumed or identified for text mining:
- Text categorization: This is putting textual terms into similar groups or classes.
- Text clustering: Here you outline a set of rules or characteristics for grouped (clustered) words.
- Sentiment analysis: This is similar to text clustering. With sentiment analysis, you presume a feeling or emotion within the text.
- Summarization: This derives a condensed statement or indicator from the entire document or block of text.
- Named entities: Here, text mining considers a relationship between a name and accompanying text in the source.
Depending on your goals and the metrics you're using to track them, you may or may not be interested in every type of text mining, but know that any social CRM solution with text mining capabilities can offer multiple types of text sorting.
In this next section, we talk about how to leverage text mining to generate useful business insights. With text mining, you close the gap where quantitative data leaves businesses wanting a fuller view of analytics.
Using data to enhance customer interaction
Knowing how and why customers interact with your brand is vital to fostering long-term relationships. You need to stay up to date on the trends among your customers in order to maintain engaging conversations with them. So, if you have 2,000 Facebook fans, you'll want to stay on top of what they're saying, whether they're commenting, and whether they're sharing you content.
Analyzing the content that's produced through social media channels can help you answer these questions accurately. Instead of acting on a hunch or on what you think is happening, you can use analytics to get real data from numbers and text. (Look at Chapter 16 for some social CRM solutions to check out.) Then you can prepare for more-targeted interactions with your customers.
Determining what metrics matter for social CRM
The central theme for social CRM is customer loyalty, but identifying it can present a challenge for many marketers. Just about everyone can agree to what customer loyalty is: a strong affinity for a brand that results in a desired customer behavior. Where the challenge exists for marketers is defining that desired end result, which is customer loyalty.
As we mention in this chapter, you'll first need to identify what you want to know about customer loyalty — which metrics — before you can start to extract data to measure. Behaviors beyond transactions and purchases can indicate customer loyalty. Perhaps a loyal customer will share your Facebook post or mention your brand in a post of her own. That can indicate loyalty as well and be rewarded as such.
Your data can help you reward customers for loyalty if you segment your customers based on level of affinity. There are many ways to look at your customers based on your brand's determination of value of loyalty. Here are a few ideas on segmentation using social CRM:
- Likely to recommend your services: Look for fans or followers who frequently post positively about you, or share your messages.
- Likely to repeat purchases: Combine traditional CRM sales information with social information about their habits and lifestyle.
- Likely to purchase additional products or services: Gather information on life events, such as marriage and new babies, as well as demographics, to predict future behavior.
- View your brand to be superior: Use text mining to track positive posts, as well as sharing behavior. You can also look at other brands a customer is connected to (or not) in social media.
- Wouldn't ever purchase a similar product from a different brand: Identify so-called superfans through text mining and volume of posts about you. Sometimes these customers even mention your company in their bio information or handle, such as MacFan4Life.
- Communicates actively with the brand: Tally posts that mention your brand, and social media conversations with your social media representatives, to find your most vocal customers.
How you measure and segment your customer loyalty can be based on short- or long-term goals. Just keep in mind that outside factors may impact some of the situations outlined previously, and they don't always correlate to customer loyalty. For example, an increase in a certain behavior could be the result of a specific promotion. Customers may stick with your brand because it's too difficult to switch, rather than because they like you. Ask yourself, are there really other options for them? Utilities especially are often faced with this situation, as customers may not be loyal to their cable or power company but they don't have many other options. Be sure to look at customer loyalty beyond transactional numbers and consider the layers of customer loyalty.
Measuring the importance of advocacy
Value exists in brand advocates, particularly in the social realm. Brand advocates are the customers who tout your brand to their network. In this way, the power of one customer can exceed their single purchase as they share their views with many people online — more than they could reach without the help of social networks. This is great for businesses. Even better? Social media enables us to the see the reach of our customers. The very public characteristics of online social networking opens us up to insights we never had before about our customers’ spheres of influence.
We need to consider the emotional relationship customers have with our brand and how they demonstrate that — directly with us and also to their network. Are they likely to recommend our brands, or are they actually going out and actively recommending our brands right now? Tapping into text mining can reveal the true sentiment of our customers to determine the power of advocacy.
Realizing the Net Promoter Score
Net Promoter Score (NPS) is a metric (and registered trademark of Bain and Satmetrix) for customer loyalty, specifically evaluating the likelihood of a customer to recommend your brand. It's widely adopted by many Fortune 1000 businesses but can be a powerful measuring tool for small- to mediumsized businesses as well. The score identifies customers in these three categories based on asking a simple survey question: How likely are you to recommend this brand?
- Promoters
- Passives
- Detractors
Businesses use the results of the Net Promoter Score to direct (and measure) employee and company interactions with customers. With insights into this metric, you can identify customer service issues and see where you aren't delivering your desired brand experience. You can then adjust your communications to meet the needs of your customers. For example, if a detractor shares negative comments about your brand, you would address that person much differently than you would your best customers: your promoters.
An official NPS score can be measured only using the NPS proprietary system, but that doesn't mean you can't approximate by looking at social CRM data or designing your own survey. Use text mining to see who often speaks positively about your brand, and connect with those people to keep their positive conversations going. Identify those who post negative comments and see how you can fix the issues they saw, and turn them around. And finally, brainstorm ways to turn passive customer conversations into positive ones.
Finding a Social CRM System to Meet Your Needs
Now that you know how you can turn social data into metrics, you can begin to determine what information is most important to you now, and think about what may be important in the future. Knowing all the capabilities of social CRM systems will open your eyes to the vast amount of customer information available to you.
Think about what you want to know, and what you'll do with that information. Are you looking, for instance, to create brand advocates, get feedback on new products, or compare against competitors? Knowing what you want will help you find the social CRM solution that's the right fit for your company.
But how to choose? There are many social CRM systems available today, and more will likely be added as social media use continues to increase. In Chapter 16, we identify the top ten enterprise-level social CRM solutions you should look into, but don't stop there. We encourage you to do your own research as well to determine the best system for your business, your employees, and your customers. What works for one company may not be the best solution for a similar company in the same industry. Look for free trials and compare various solutions on data collection, reporting capabilities, and user experience.
Analyzing the Future of Analytics
Marketing and PR initiatives have their own set of metrics. What works for those avenues toward customer engagement can't be applied with a blanket over social media. Each tool used to reach and engage customers requires its own set of metrics.
To make social media analytics deliver the metrics you need, you need to define your standards for measurement. You can build an analytics framework that you can apply to various social media platforms. What works for your industry may not work for the next. Here are two metrics that you can use across multiple channels and for many industries:
Capturing a deeper understanding of our customers in our social CRM systems is gold. With predetermined standards for measurement and the ability to segment affinity levels, you have more than sufficient knowledge to adjust your business's communications to meet social customers when, where, and how they want. Remember that social CRM aims to enhance customer experiences by creating a corporate culture that fosters customer-centricity in every facet. This has to be true as you look for metrics and goals and analyze customer interactions.