Everyone who has ever worked in Data or Social Listening and been involved in building queries will hopefully relate to what I am about to write.
The first step in all the briefs we receive is to come up with a query that will enable us to retrieve all the data we need to surface insights. Even if, initially, writing queries doesn’t seem like a complicated process, it is! It is a full time job that requires a very specific skill set.
An art and science
At Black Swan Data, we often say building queries is an art as well as a science, which balances a lot of parameters forcing us to face key questions like:
What shall I add to my query? What shall I remove? How many Keywords will I use? How many operators? If I include that keyword, how will it impact the full dataset? What if I include a keyword that will bias everything? Do I need brand mentions or shall I focus the analysis on the core keyword?
But whyis it more challenging?
Because the brief is more complex? Because the scope of work is larger? Because their brand ecosystem is vaster?
No, the answer is so simple and obvious that at first you may not even consider it.
- Sky doesn’t mean internet provider.
- Sky doesn’t only refer to mobile, TV offers and broadband plans.
- Sky is not, per se, a brand.
The Oxford dictionary defines Sky as:
“the region of the atmosphere and outer space seen from the earth.”
And that’s it, all of a sudden the already complicated task turns into a chimera.
How am I supposed to build a query when the name of the brand also refers to birds and clouds? How will I find a way to target the brand mentions but avoid downloading all those inspirational quotes people love to share on Pinterest, like
The Sky is the Limit?
But we love challenges at Black Swan Data, and we love turning what sounds impossible into an amazing piece of insight.
Desk research
Every data project we undertake here is preceded by a desk research phase that enables us to develop a thorough understanding of each market and social conversation.
Using different tools developed by our team of data scientists and developers, we start by observing social conversations and identifying the core elements being discussed. We then advise on how best to structure the query based on the specifics of your brand. This allows us to develop clever ways to retrieve brand specific data, rather than simply going for a volume based approach.
By juggling with operators, parenthesis and additional keywords, we come up with the best possible query to get the most accurate view of the conversation about your business.
Minimising noise
On top of this tailoring phase, we also spend quite a lot of time doing additional research to identify keywords causing noise, embedding exclusion lists within queries to minimise the amount of noise and give you more accurate data.
We often hear in big data that everything is about quantity, but with the previous example we can say that if quantity comes first, quality is definitely not far away. This is even more true when your brand is a homophone, and you are trying to get an accurate view of your share of the voice.
Today I asked around the office for names of brands that are also nouns. In just a few minutes we identified several:
APPLE, AMAZON, VIRGIN, SKY, BURGER KING, JAGUAR, SHELL, ORANGE, DOVE, BOOTS, ICELAND, MARYLAND, MAGNUM, INNOCENT, WALKERS, SALT, CURRIES, MANGO, GUMTREE, GUESS, WINDOWS, VICE, VANS, OFFICE, FAY, QUAKER, WALKERS, WASABI, COKE, WAREHOUSE, O2, ROGERS, WIND, THREE, RED BULL, MARVEL, BLACKBERRY, HABITAT, VIEW, ALPHABET, SUN, MIRROR, MARS, BOSS, VISA, JOY, TIDE, PIZZA HUT, CANON, SPRINT, OMEGA, STAPLES, TACO BELL, MALIBU, CHROME, TARGET, ARIEL, FLASH, ADOBE, ORBIT, EXTRA, HOLLYWOOD, ELEMENT, ECONOMIST, OYSTER, ORACLE, PUMA, PRINCE, CATERPILLAR, GLOBE, PERISCOPE, PINK, FAT FACE, BULL DOG, OXFORD, PANDORA, SWAG, MINI, NEST, HIVE, CANARY, HUE, LLOYDS, DIESEL, WRANGLER, DRUMSTICK, STEAM, VALVE, FALLOUT, BLACKBOX, SLACK, VINE, SLEEP CYCLE, SUBWAY, FEVER, DOJO, MARIO, UBER, WISH, FANCY, AMERICAN APPAREL, COMFORT, FAIRY, PEDIGREE, PLAYBOY, SAFARI, MINUTE MAID, TIGER, GIRAFFE, NEW BALANCE, CAMPER, GALAXY, MILKY WAY, TWISTER, NEXT, MONOPOLY, SHARP, CHEVRON, MONSTER, CORAL, UPS, NATIONWIDE, SMARTIES, DOMINO’S, JIGSAW, SAW, FORD, ASPIRE, WALLS, SHARD, ROUGE…
So the challenge is real but not impossible. And this is what we do and what we are proud of at Black Swan Data.
