Metrics are the Holy Grail for internal communicators. They can extract insights to engage with employees across the whole organisation. Read data carefully and it becomes something very useful—a guide on how to navigate the business.
Measurement can provide practitioners with a ladder to meaningful decision-making: what the best channels to interact with colleagues are; what the best content is; what audiences to reach at any given location, time and context; what messages should be reinforced; what stories should be curated that appeal the most to staff.
But, one of the main challenges in today’s digital and social networking age is to catch up fast with all the amount of information flowing inside the company. Which raises a pressing question: how can practitioners maximise the power of measurement to support the creation of relevant communications?
Even the most initiated need to learn how to game the system. An important starting point is to look at who is doing it well. No doubts that Marie Wallace (pictured), Analytics Strategist at IBM is one of them.
I wanted to speak with Wallace to explore how internal communicators can exploit data from enterprise social networks (ESNs) to benefit both individual staff members and the company in its entirety. In this interview she shares her view on the state of measurement, the move from descriptive to predictive analytics, the power of the social graph, and the partnership that data scientists and communicators need.
Gloria Lombardi: As part of your daily job, you interpret data from the Social Graph of companies to help them make better decisions and communicate more effectively.
As for those who have not your same level of understanding, could you explain what do you mean by Social Graph, and how evaluating it can help communicators?
Marie Wallace: The Social Graph is the representation of how people interact inside the organisation. I also call it Enterprise Graph. It takes into consideration all the different communications that happens inside the company. This includes all the direct interactions such us having a telephone call with somebody, or sharing a text via instant messaging (IM) or having a discussion on the internal platform.
It also includes all the indirect connections. For example, I create a presentation around a certain topic; then some people download that document; they recommend it to other colleagues who may also share it with somebody else; others may start liking it and making comments on it. All those communications happening with me because of that document are considered indirect, and can be all tracked.
GL: What does this mean for communicators? How can they make the most of it?
MW: When you look at the whole business and all the different ways people communicate with each other – directly and indirectly, you can piece together a big network.
Ultimately, when you analyse the Social Graph you are trying to understand how staff interact with each other inside the company to get their job done.
Once you put that graph together, you we can start analysing the organisation, asking a whole set of questions from collaboration, to engagement, innovation and retention for example.
GL: Based on your extensive analysis with a variety of businesses, what is the current state of analytics? What is emerging?
MW: I am seeing a couple of fundamental changes. Until recently, internal communicators were the people responsible for getting the message out. They were crafting the content and spread it to A, B or C. They were the owners of the message. They had the control.
With social media – both at enterprise level and external networking channels – the message has a life of its own. In fact, what is becoming increasingly important for communicators is their ability to get the community spread the message – not for them owning it.
It is challenging for them as they have less control.
GL: As a result of them having less control and the community gaining more power, what do you think is important for communicators to understand and do?
MW: It is increasingly important that they understand two things. First, to understand how information flows inside the company: What are the right channels, mechanisms, times, types of media to allow the message become alive? Certain communities might make a lot of use of videos; others may prefer blogging. Some employees may want to read news during the day; other will check it early in the morning. As a communicator you need to understand well all these pieces of information.
Secondly, because they are now relying on the community, they need to understand not just what the most pertinent channels, media and times are, but also who the best people are to help them spread the message and ensure it resonates with the broader community.
GL: You are a scientist with a broad knowledge of data. How can internal communicators, who do not necessarily have the same level of familiarity, interpret the Social Graph in a relevant way?
MW: They need partnerships with analysts. No line of business is going to be able to single-handedly figure all this up.
As a data scientist one of the first things that I have to do with all the analysis that I do is to create a visualisation of the Social Graph in such a way that anyone can understand.
For example, when we crated the dashboard for our own employees at IBM, we never talked about things like graph analysis. It just gave and gives our colleagues the measurement on things that are relevant for them – how do they compare with their peers; how much action are they generating; how much eminence are they generating from their interactions; how well respected they are within their networks and so forth.
GL: We need these analysis solutions to be built for internal communicators.
MW: Yes. Practitioners don’t need to know or worry about deep analytics. They need the right recommendations for the best channels to use, the right people to contact to build a network of advocates, etc.
Data scientists can create this information through simple interfaces that can make communicators feel very comfortable with.
Perhaps the market is not ready yet for those solutions. But this is absolutely where we have to go. It is not just communicators who need it – any line of business doesn’t want complicated data analysis. The only people who like analytics are analytics people like me! The rest of the organisation wants their questions answered.
GL: Recently, there has been a lot of talk around the move from descriptive to predictive and prescriptive analytics. Could you give me your view?
MW: Generally speaking, descriptive analytics describe what happened today and what happened in the past. In contrast, predictive analytics predict what is going to happen in the future or what potentially could happen in the future. Prescriptive analytics is the most interesting piece for me because it tells you what you should do – there is the actionability piece.
We have been using descriptive analytics for decades with transactional data. For example, measuring how many deals the sales team closed in Ireland during the last three months. It tells you a lot about what has already happened.
The challenge in the social and collaboration space is that we look at people and networks, unstructured data, which traditionally has not had the same level of descriptive analytics. Because of that, we still have to do a lot of work to answer simple questions – with big data we have massive volumes of information that describe what is happening today and what happened in the past. Even inside a small company there can be hundreds of nodes and edges, if you think of every interaction that happens among individuals, content and business processes. Yet, most companies don’t measure what is going on.
I am absolutely convinced that we are moving towards predictive analytics. This is hugely powerful. But first of all we have to be able to answer the basic questions – What is happening in our system today? If we can answer that, we will be more likely able to predict what is happening in the future.
GL: What tools is IBM using to capture and evaluate all those different data sets inside an enterprise?
MW: We combine open source technologies with IBM proprietary intellectual capital around analytics. Working with open source technology allows us to represent the network in different ways. On top of that we use our own decades of research around network analysis and technologies like IBM Watson, which helps us to extract deep insights from content and use those insights when we analyse the network. Plus, the IBM predictive analytics package and the platform for IBM biggest insights.
GL: How can analytics help organisations appreciate the value of individuals and the importance of their networks?
MW: There are many different ways to analyse collaboration and social data to help both the individual within the company and the company.
For example, at IBM we have been working heavily around the ‘democratisation of analytics’, by focusing on putting data into the hands of the user. We understand that in order to improve the adoption of social within the business, we have to provide individuals with the information that helps them understand how they are doing. So, for example we help them recognise if the content they create resonates with the community; if and how their colleagues appreciate and interact with that content, what is working in terms of channels and what is not (e.g. something they share via IM receives a small amount of interactions, while blogging gravitates the conversation).
Another thing. From the social graph analysis you can combine high correlations between engagement at organisational and individual level, as well as the likelihood of innovating. All of this can be very interesting for communicators.
GL: You mention big data and the high volume of information that our organisations are facing today. How should communicators approach and condense that data into meaningful actions?
MW: Don’t try to boil the ocean. Don’t look at everything at the same time.
One of the biggest challenges that we have is that people don’t even collect data. Or if they do collect data, they do not use it.
Another important thing to emphasise is that there are many different questions that you can ask of the data. First, think at what questions you want to ask. Secondly, look at where that data resides. Then, start analysing it.
For example, if you are at a company where there is a major re-organisation your biggest interest should be around understanding how the network is currently structured. If you are thinking of moving away somebody from a particular unit but it turns up that they are highly influential there, pulling them out would cause a collapse of the network. You may prefer finding an alternative solution.
But, if your interest is looking for expertise in a particular field, you would need to look at another type of data set, which tells you who the people with that knowledge are and how they interact inside the organisation.