Big Data Buying Criteria for Social Engagement
By Edwin Margulies
I encourage those shopping for a professional social engagement for customer care platform to consider the Big Data elements of this decision.
Why? Because professional-grade social engagement platforms are built on a foundation of Big Data.A modern and effective platform needs to be able to capture, curate, store, manipulate and help you visualize and act on essential data. In this article we explore some of the essentials that are often taken for granted in selecting your winning platform.
What's Big Data and Why Should I Care?
Big Data is all about assimilating and making some sense out of huge amounts of data and then making intelligent decisions about it. The challenge presented by Big Data is summed-up in John Naisbitt's (Megatrends) 1982 futuristic quote: "We are drowning in information but starved for knowledge." So the big challenge Big Data represents has been with us for a while. It's just gotten bigger. In fact, this year a Forbes article quoted IBM as saying that 90% of all stored data was generated in the last two years. The reason you should care in the context of social engagement for customer care? Because what people say, where they say it, and how it affects your brand also gets bigger every day.
Consider this. In one work week, Twitter logs over a billion tweets. In one month, 70 billion pieces of content will be shared on Facebook by over a billion members. And there are 92 billion YouTube page views every month. The list goes on, but you get the idea.
What are the Essential Elements of Big Data?
For the purposes of shopping for a professional-grade social engagement for customer care platform, you can look at Big Data as forcing a series of tasks. These tasks need to be performed in order to deliver useful information (and ultimately knowledge) that you can use. In your company, the users will be social care agents, supervisors, business analysts, and managers. The essential elements to consider are:
1. Raw Data Capture
2. Digital Curation and Management
3. Data Dissemination
4. Analysis and Visualization
Raw Data Capture
You may take for granted how easy it is to lay your hands on the raw data. But any world-class social engagement platform will need to have some kind of a "firehose" data feed. The leading platforms handle this with arrangements with data feed providers such as DataSift and GNIP, for example. These companies offer access to raw feeds from Twitter, Reddit, Blogs and Articlesto to name a few common sources.
But it goes deeper than that. For example, if you have a corporate LinkedIn page, you can't get that fed to you by a regular aggregator. You need a platform that has native API integration to LinkedIn for that. Ditto Facebook fan pages. The ability to respond and reply to disparate fan pages in a consolidated platform cannot be taken for granted. How about links to "private" community sites and forums? That does not get delivered via a fire hose vendor.
In other words, modern social engagement for customer care platforms must go beyond the "usual suspects" of fire hose aggregation and provide the flexible ability to add on "private cloud" data streams. This is one of the chief tenants of Big Data - the idea that the data can come from anywhere. The more anywheres your platform can deal with, the more relevant data is available to then curate and manage. Therefore, when making buying decisions, remember to ask questions about the flexibility the platform has in grabbing data from virtually anywhere. Ask if custom listening engines can be set up to listen to private data clouds. If you do not ask these questions, you may end up with limitations that could hurt your engagement initiative.
Digital Curation and Management
Curation is the discipline of taking digital assets and collecting them into meaningful data sets and then preserving the data. This preservation must be done in such a way that it is accessible and maintainable. In social engagement, this requires the use of curative languages and Boolean math to set up search parameters and arguments so you can sift through large data sets and pick out the data you need. For example, if you run a grocery store chain, you might want to set up search parameters that allow you to pick out "East coast mentions" of a big sale you ran on XYZ brand chicken breasts. Perhaps you ran out of product or there was spoilage and a lot of people complained. Regardless of the customer service ramifications, your ability to zero-in on specific data sets out of an ocean of data - is part of the curation and management aspects.
Further to this basic search capability is the need to automatically cluster and package information that is trending. To do this you need to go beyond basic filtering and use real Natural Language Processing engines. If the platform you are looking at does not have the ability to automatically bin posts into business issue clusters or trending topics, the basic curation and management of the data is lackluster. This discipline of curating the data along with artificial intelligence is not to be taken for granted. You actually have to ask: "Does this platform automatically understand spam, sentiment, clusters and trending topics?" If the answer is "no" the curation and management of Big Data is not robust and you'll end up being disappointed with the results.
In the context of social engagement for customer care, dissemination is all about arranging meaningful data for transfer to your data constituency. These are agents, supervisors, and business analysts and managers that use the information.
The acid test for robust dissemination is to make sure you are able to discretely search for, transfer, and share required information. If your platform can't do these things effectively, it's falling down on the job Big Data-wise.
Here are some questions you should be asking about data dissemination: "What is the mechanism for searching on specific author data, drilling down into specific topics and linking outreach data to that author's posts?" Also: "How are topics and clusters filtered or routed to certain agents based on their workgroup or skill?" And; "How can I sift through all the data to route only angry customers to properly skilled care specialists"?
You should also consider how decisioning figures in to Big Data. You should be able to link Big Data elements and trigger certain actions on that data. For example, you should be able to route and escalate a social post if it: 1) belongs to a cluster representing "at risk" customers; 2) is from an author that has a public influence score over 20%; 3) is from an author who has a corporate influence score of over 30%; and 4) a person who is not happy sentiment-wise. You should ask if automatic routing and dispositioning can be done based on these elements.
Analysis and Visualization
A big goal of leveraging Big Data is to take relevant action based on knowledge gained from the data. This is only possible if the data is packaged in such a way that it makes it easier for you to view this mashed-up data and then make decisions on it. To put this into the proper context for customer care - be sure to look for role-based reporting that makes sense for your operation. Ask: "Does this platform serve up KPI (Key Performance Indicator) data on an agent-by-agent basis?" And: "Can agents see a self-KPI view based on personal performance?" And: "Can I visualize word clouds and trending topics so I can make real time decisions on outreach and dispositions?"
Another thing you should look for are rich visualizations that provide instant insights into customer and agent behavior. For example, a multi-colored graduated bar chart signifying sentiment ranges and agent interaction can be helpful. Imagine being able to spy a dashboard showing agents' level of outreach with angry customers vs. happy ones.
In short, the most important consideration with Big Data analysis and visualization for your engagement initiative is usability. That is to say, your ability to actually use the data that is presented and put it to work right away. This reaches in to visualizations that help you to coach agents and visualizations that instantly identify hot spots - all the way to trend analysis. Be sure to see how all of these work even if it is acting on sample data.
When considering your choices for a professional social engagement platform, it is important to consider the essential elements of Big Data. How adaptable is the platform in its ability to capture raw data? How powerful is its curation and management capability so it can wrestle the important trends out of large hunks of data? Can the platform "slice and dice" the information and easily serve it up to the right people? And can the data be transformed into usable knowledge so you can visualize and then take action on it for your customers? These Big Data oriented questions should be top-of-mind when you go shopping for a social engagement for customer care platform.