The search for that killer insight
Data, data, data….for some it is the lifeblood of their business with decisions being driven by a complex morass of probabilities and analysis. for others, data is a magical, mystical vortex of mystery both intriguing and scary at the same time. They know worth is there, but just don’t know clearly how to leverage it.
For others, there is inherent mis-trust, based on the strength of gut assumptions over many years of making decisions and largely getting them right:
“People living regionally don’t use computers do they?? So we must send them post and only ever post.”
“Older folk don’t use social media!! My Gran would have no idea what an iPad is!! We must advertise in the paper!”
However you use data, or perceive the value of it, for a number of years terms like ‘Big Data’ have dominated many products, media coverage and conferences. Data continues to be a focus for many businesses looking to get, or stay ahead.
We try to take a data informed approach. We seek many data sources, make sense of them and then overlay with a human factor to make the decisions we need (this in itself was inspired by an Atlassian speaker at UX Australia 2016).
Of late, we’ve been trialling many tools and implementing some that can help with this objective. Two of these are Outlier and Friendly Data.
Outlier
As the name suggests, Outlier analyses many data sources in order to surface the outliers in your data. Data sources can range from the easier to access SaaS and social nework sources to the sometimes hideously complex beasties such as ERPs and CRMs. Outlier work with a business to help set up this access before the service starts.
Once you get up and running, the results you receive are a mix of automated alerts and items vetted by their account team — both aiming to bring to the client’s attention the anomalies in their operations. We initially tried using a mix of Google, Facebook and ERP data but we quickly learned that trialling with static data sets isn’t what this service is about. The strength is feeding in data as often as possible, and is future focused — you can’t dump in 10 years of sales data, it needs to be analysed on entry. The longer it analyses, the better it gets at finding stuff.
After our initial set back, we continue to run the service albeit with only Google and Facebook data. Google increasingly is providing similar kinds of reporting and alerts, so we’ll see if these outpace the service Outlier provides. Of course, the enrichment of data through multiple sources is the strength of Outlier — so we’ll continue for now and keep fingers crossed for that killer piece of insight.
Friendly Data
This is an interesting one, and we haven’t seen many other services like it. There are definitely some, and even Power BI partially offers this functionality. But we were intrigued by a service that offered natural language processing to query data sources and so we got in touch with Friendly Data to organise a trial.
As with Outlier, we worked with the Friendly Data team and supplied them with a subset of our anonymised data to set up the service and to get a feel for the results. The trial version didn’t include a GUI which was limiting with regards to the group that we could trial it with but we gave it ago. Initially got some promising results but also saw some holes in the set up. Our query “Return all the customers under 40 years of age in Clarkson” returned zero results. After working with Friendly Data, we discovered the query “Return all customers under 40 years of age in Suburb Clarkson” would have yielded the results we expected. So not quite natural language yet.
Nevertheless, a promising service and Friendly Data do have some exciting updates planned which we’ll look further into.
