Last week I had the opportunity to present my views on social media analytics at Business Insider’s Social Media Analytics conference (you can view the slides from my presentation here). The lessons I presented were based on my own experience, both based on my work at SAP but also on discussions with my peers in the industry. They were centered around the following three key takeaways:
1. Differentiate, but be consistent
Although this sounds like an oxymoron, defining a consistent social media analytics framework can help create a common vocabulary and allow you to learn by being able to consistently compare across your various campaigns. There are three key dimensions as shown in the chart above:
- Use case: It starts with understanding what you are trying to accomplish and then driving your strategy and results based on those objectives. Are you trying to build awareness with an audience not necessarily familiar with your brand and offerings, or are you trying to generate demand running a social media campaign? These are two very different things and the way you measure success needs to vary based on these objectives.
- Stage in the audience’s journey: This refers to the stage of the journey your audience is: are you building a large audience that is engaged, doing the things you would like them to do, and encouraging ongoing loyalty and advocacy? I do recognize that it looks surprisingly similar to a standard marketing funnel, but that is exactly the point.
- Social media channel: While the least important dimension, you do need to keep this in mind, as how you measure each of the previous two dimensions will vary across the various social media channels.
2. Don’t try to outsmart yourself
This is an area I see as a very common trap we always seem to fall into, especially as we get started. My key premise is that you have to ensure that your social media analytics are in lock-step with your overall maturity, which I like to call Social IQ. This translates into the following questions:
- Are you asking the right questions, or are you ‘boiling the ocean’? Do you have a hypothesis in mind that you are testing against?
- Do the volume and depth of social media analytics reflect your maturity level? If your organization is unable to consume a ‘level 0’ metric, what is the point of going to ‘level 7’?
- Are your social media analytics too fancy or incomprehensible? One of my favorite examples is the ‘social media engagement index’: Add your Facebook likes and Twitter retweets, and divide them by the sum of your Facebook fans and Twitter followers. How can you ever make sense of such a figure that attempts to combine apples and oranges?
3. Invest wisely
Once you have laid down the basics, there are four points to keep track of as you scale:
- Think use cases, not social media analytics tools: Think through the problem you are trying to solve first before building or buying tools. Make sure you tie whatever you do to business outcomes as I mentioned earlier and don’t treat social as something on top.
- Don’t be afraid to experiment: This is tightly linked to the point above, and you have to try relentlessly. You will fail, but the point is to fail fast and forward, and learn from your mistakes before trying to automate anything. The key here is to start small first, learn from your experiments, and grow from there.
- Pick partners wisely: Although many social media analytics vendors claim they can do everything today, don’t believe them. This space is still in its infancy and although consolidation has already started, it will take a few more years for the space to mature.
- It’s OK to outsource, but build a plan to in-source: It is very natural (and even advisable) to recruit consultants and agencies as you embark on your social media efforts since you will likely not have the right skills and resources in house. However, you need to both think hard about what to outsource (you can read my point-of-view here), and prepare to in-source or automate tasks as soon as you start to build scale – what is the point of hiring agencies to compile data or perform basic business intelligence tasks when there are many tools that can do this today?
What do you think? I would love to build on this based on your own experiences.