Let’s Talk About Social Analytics - 2017 Edition

I gave a talk at Social Media Week Copenhagen (#SMWCPH / #SMWDK) a couple weeks ago, and it was a much needed update to my “Let’s Talk About Social Analytics” post from last year. While not much has changed, the urgency of what I talked about is the reason why I had the session.

What’s changed then? Here’s what.

Too much data, not enough analysts

I did the usual thing of asking the attendees to raise their hands if they were (1) community managers (lots of them), (2) social media managers (not as many, but still lots), and (3) analysts in any capacity - social media analysts, digital analyst, web analysts, marketing analysts etc. These were in the minority: in a room of roughly 90 professionals attending a session I listed as “by analysts, for analysts”, there were no more than 8 analysts in attendance.

Everyone agreed that data analysis is an increasingly important demand that needs to be fulfilled, yet not a lot of these companies are doing much (if anything) to fill that gap. There’s an exponential increase in data being generated, in data that needs to be measured, in data-driven decisions that need to be made, decisions that are still, somehow, being made by gut, feeling, tradition (“this is how we’ve always done it, so this is how we’re always going to do it”), or by the opinion of the highest earning person at the table (this seems to be the norm in bigger companies).

Get yourself an analyst. Whether you want to have this role in-house or outsource it via an agency, whether you want to hire someone new or you want to find someone who already works in your team/department/company with that analytical inclination, whether you want to call this person a social analyst or digital analyst or online analyst or whatever other title you think fits best - how you want to have this role is up to you, but the need for an analyst shouldn’t really be up for debate. Not in 2017.

Get the right tools

I’ll reserve a separate post specifically on tools, especially on the new tools that have popped up in the last 12 months, but there is one thing you need to bear in mind: whether you choose to hire a social analyst or you already have one (or, why not, more than one), the analyst needs tools to do the job. Not just any tool - the right tools. Sure, there are so many social tools available today, but not all of them are worth your time. The one thing that does not exist is the perfect social analytics tool, despite what a few social vendors may tell you. What does exist, however, is a set of tools that fits perfectly within your social analytics requirements, thus your business requirements.

Business Intelligence (BI) tools will be especially important this year onwards, as more social analysts start coding to get their data - more on this later.

Democratise social analytics

One of the traits of being a social brand (slides 9 to 12 of my presentation) is a change in perception of social analytics: (1) elevating it from just social analytics to social intelligence, and (2) democratising social analytics. While I covered the first bit during my session, the second portion does need some clarification.

Gone are the days of analysts reporting one-to-one, i.e. having one stakeholder request a report, the analyst compiling the report solely for this stakeholder, then moving on to the next request. This workflow, although common, reinforces the perception that analysts are just “report churners”. Like I said in my session, we’re much more than exporters of Excel spreadsheets and creators of PowerPoint reports. In other words, we’re much more than what advanced tools can do on their own. One of the things that sets us apart from analytics tools is the ability to layer social analysis from multiple sources (including sources that aren’t tied to one tool, e.g. sales from the last quarter, NPS from the last financial year, results from an opinion panel), and being able to supply intelligence from knowledge, experience and hindsight - something that a super-intelligent tool like IBM Watson wouldn’t be able to give you, for example.

So, what do I mean by “democratise social analytics”? Every time you get asked to run a report or supply some analysis, stop and ask yourself - who else could benefit from this? What are the implications from my findings? Who’s likely to be responsible for the actions and recommendations I supply as part of this analysis? Most reports you run should be one-to-many, rather that one-to-one: find the “many” beyond the one person who asked you for the report. Make your reports accessible to everyone in the company. Use privacy settings and your judgment for any sensitive reports that not everyone should be able to access (most file repositories, like SharePoint, Dropbox, and G Suite will let you tweak privacy settings accordingly, so you don’t have to worry about unauthorised eyes seeing sensitive reports). Make sure that social isn’t a trade of the few but a responsibility of the many (this is where social advocacy comes into place). That way, everyone knows how social media can impact pretty much every part of the business - sales, marketing, the team developing your web assets (including your website and apps), HR, PR, your C-Suite, and beyond.

Time to code

Let’s be honest - good tools are becoming scarce nowadays. Correction: good tools at a reasonable price are becoming scarce nowadays. While “reasonable price” is very subjective, we do have too many tools charging extortionate amounts for a mediocre platform, and this is especially the case for social suites. Social suites are great for joining multiple social responsibilities in one single platform that everyone has access to, but when it comes to analytics, solely relying on a tool that doesn’t focus on social analysis is an awful decision.

I mentioned that 2017 is the year we see more social analysts evolving into data scientists, into data analysts that collect data from more sources than just social, into data analysts that are more self-sufficient. To get ready for that evolution, social analysts of today will have to get to grips with programming, specifically two languages - Python and R:

  • Python is ideal for grabbing publicly available data from the web. You can either use python to access APIs or to get data off a webpage (“scraping”). Thankfully, a lot of social platforms make their APIs available, including Twitter, Facebook, YouTube, Instagram, and Reddit.
  • R is built with statisticians in mind, so it’s perfect for data analysis, statistics, data models, and especially data visualisations. R lets you visualise data in a more effective and efficient way than rows of numbers in spreadsheets.

If you can master both Python and R, you’ll eventually find yourself relying less and less on 3rd party social analytics tools, which will in turn help you save the money that you’d normally be spending on tools. This isn’t to say that Python and R will completely replace tools: BI tools are still going to be important for analysts, if not more. But, if you’re already using a poor or mediocre social analytics tool (including social listening, social monitoring, social management tools, and social suites), mastering Python and R will give you the right tools you need to do the job.

More social analytics tools are picking up on this trend: for instance, Brandwatch recently released their second iteration of their visualisation platform Vizia, this time going beyond just social data visualisation, with “a new open development framework, powerful new visualizations and intelligent new ways to tell the business stories that matter to you”. Then you have other BI tools like Power BI that are constantly beefing up their data analytics and reporting capabilities (too many for me to list here, but you can check out the Power BI blog for more info). For instance, I sometimes use Power BI to correlate my social activity (mostly on Twitter) with the traffic this blog gets (which I mostly track via Google Analytics). While Google Analytics has some social capabilities, it is very, very basic, and it’s nowhere near the amount of data you can get via Twitter Analytics…so why not join both? Joining Google Analytics (an expert in measuring web traffic and conversions) with Twitter Analytics (the expert in measuring Twitter analytics) via Power BI (the expert in business intelligence and analytics platforms).

You’re not alone

One of the things I mentioned towards the end of my session was that, just because there’s not enough of us, and just because social analysts are far and far in between, it doesn’t mean you’re alone. Find yourself a mentor, especially if you don’t know how to get started, and if you don’t know who to ask for advice or some mentoring, I’m available; find yourself a group of people that you can rely on for help with any social analytics dilemmas. One example is the Social Media Analytics group for Quintly (specifically built for social media analyst and social analytics enthusiasts in and around London), which you can join here. There might not be many of us, but you’re definitely not alone.

Anyway, if you want access to the slides, here’s the link to them.

As always, if you have any questions, just give me a shout.

MarTech List 200

Google's SyntaxNet and Sentiment Analysis