Interview: Exploring Data Science at Brandwatch with Hamish Morgan
By Olivia SwainAug 23
The days of simply plugging in your brand name and trusting the data behind the line graphs in your SMM tool are over.
CEOs and CMOs are becoming increasingly social and if it hasn’t already happened, then soon the day will come when they call you into their office to defend how you are measuring “social”.
The scary thing is that social data – and the Internet in general – is a bit of a hot mess, overflowing with constant inputs from imperfect humans and trolling spammers.
Setting up a query to track the volume of both your brand and your competitors is no trivial matter, and ideally should be something your community manager, brand manager, marketer and even the CMO and CEO (or your agency!) should be weighing in on.
Why? Because how you create a query can change the results of your measurement in a huge way!
At the highest level the quality of your query results can be affected by:
A lack of measurement strategy from the top down
Without a comprehensive, well-thought-out query strategy you can end up with flawed or deliberately skewed data that could be being used to make important business decisions, measure marketing ROI, or even promote or fire employees/agencies.
Creating a Query Strategy
As an enterprise, understanding what, why and howyou are measuring social activity is essential for social media based insights or research – whether you are trying to answer quantitative questions like “volume” or “share of voice” or qualitative questions like “what did people think of the super bowl commercial?”
When it comes to setting the query strategy, there isn’t a strict right or wrong, but what’s important is that companies think through things like how they define a social “mention”, and why they want to measure “volume”, for both their own brands and their competitors.
Even if the goal is to just look at volume of mentions, it is very important that the CEO/CMO understand how the query strategy will affect the results.
A few examples of things to think through at a basic level are:
The brand name itself – are you including every possible version of a brand name, including common local slang versions and misspellings?
And at a more advanced level:
Do mentions of a sale by a reseller/vendor/affiliate or even someone selling their 1999 Ford F250 count as a “social media mention” to your company?
Of course – thinking beyond simple volume tracking – how you answer the above points will also affect how reliable/relevant the insights around your brand mentions will be.
The above are just a few examples that any solid data scientist will tell you MUST be considered to get reliable social insights. You must also have a social media-monitoring tool that can achieve the query strategy you select.
If the insights team, the community manager, the brand manager, the CMO and the CEO all have a different understanding of – or answers to – the above, then what is actually being measured and how?
Very possibly key decisions across the enterprise will be made on skewed or incomplete data.
If you have never thought about your query strategy, now’s the time to start. And the most important question to ask yourself is this:
What do you want your social media monitoring results to actually tell your team?
And if I were you, I’d aim to answer this question before your CEO or client asks you.