Making Sense of Big Data: Insights from the Google Partner Summit 2018

In the digital space, we have access to reams of data in every sense. We are obsessed with capturing everything that is happening and segmenting it just in case we need to draw some insight from it later.
From knowing that anonymous user #82010243 loves to visit our website every Thursday morning at 5:45am, to them always clicking on a display banner, to them submitting a request for quote or exiting on our contact thankyou page – data allows us track every movement. We’ve got our custom content groupings, our custom channel groupings, our custom attribution models and we somehow convinced everyone we work with to spell their UTMs correctly and put them in lower case.
Now with Google’s Dataset Search entering beta and ABS datasets becoming more accessible than ever we can access 3rd party data at our fingertips as well.
At Indago we are particularly obsessed with collecting and using meaningful data, so our challenge evolves into how to squish all this information down into a palatable format for those less excited by exceeding the row limit in excel (1,048,576 rows by 16,384 columns per worksheet if you were wondering).
In September, we attended the Google Partner’s summit in Sydney, and a lot of the discussion was around harnessing big data in a meaningful and useful way, rather than just celebrating how much data we all have access to. Elizabeth Fox, a Director at Google Marketing Solutions for AUNZ talked about how Google is levering products and providing resources for others, to solve for humanities big problems. We heard a collection of inspiring case studies: how a U.S. Teenager used Tensor Flow to create an image recognition algorithm that can detect early stage breast cancer in mammograms; how Scientists in Western Australia partnered with Google to track Dugong migration patterns with AI and Machine Learning (cutting their process time from 730 hours, to 34 hours); and that you can point your phone at a lamp and it will sing you a song about it. The latter is not quite as significant a problem for humanity but, a powerful application of image recognition and natural language usage.
We are into our M.L. and A.I., taking every opportunity to leverage the amazing capabilities of intelligent computing to make our work lives easier and increase the breadth of data we can process.
But even with all the technology and data in the world, drawing insight and creating actionable items remains the key to every aspect of what we do. Enter Dave Booth from Cardinal Path. Dave always steals the show at the Partner’s Summit, with his ridiculous imaginary product marketing scenarios and his ability to talk digital marketing and Google updates without living in the rose-coloured glasses version of digital marketing that some presenters seems to. You’re guaranteed to be entertained and learn something at the same time. This year my key takeaway was Cardinal Path’s approach to analysing data and presenting insights, something that a lot of data analysis seems to miss. In that vein let’s look at some simple checkpoints that Cardinal Path use when communicating about data, that you can incorporate in your reporting to ensure you’re driving value for your clients.
Value
Does it tell you something you didn’t already know?
It seems simple, but it’s a fundamental facet of drawing insight. You can tell someone about all the exciting revelations you’ve drawn from the petabyte of data you had your Machine Learning algorithm analyse, but if it doesn’t tell them anything they don’t already know, what’s the point?
Novelty
Have you gone beyond merely sharing facts and observations?
Again, a simple concept. But it’s unlikely your client is paying you to tell them that the 500% in the ROI column means they made $5 for every $1 they spent, or that the USA had more sales than Australia. Your job is to answer the WHY. Was it the remarketing strategy that drove the higher ROI? Was it the updated region-specific ad copy that increased US Sales?
Significance
Could someone disagree with you?
We should always ask if the results and insights we are presenting are significant; how confident are we that we are correct? Think about seasonality, think about other marketing activity, think about competitor activity. Think about causation, correlation and Nic Cage (they’re 95% confident that there is a 66% correlation). If you’re not confident that your analysis is correct, then how can you expect your recommendations to be effective?

Retention
Can the audience explain it to someone else?
Just because you can explain your process and insights to your team and your manager, doesn’t mean your client can understand what you’re on about! Think about how easy it would be for your client to explain it to their CEO or to the company in the all hands-on Monday. It’s all well and good to pat yourself on the back for how clever you are, but what use is that if no one else can communicate what you’ve told them?
Impact
Does it lead to a relevant business implication?
Putting together an in-depth analysis of cheese consumption and golf course revenue might get a slight nod from your golf course client, but are they going to invest in dairy manufacturing?

What value are you delivering by way of this insight? What is the impact? This is the most important check, and if we’re honest with ourselves, a lot of the time we are quick to draw cool, interesting insights and sometimes the usefulness of the insight falls to the wayside. In a less extreme example think about how likely it is that your client can act on the data your present to them. If you are presenting a massive data piece that highlights the correlation between mobile sitespeed and paid search quality scores, but they have no capability to improve their sitespeed, what is the impact of the analysis?
If you’re obsessed with data, these 5 checkpoints should help keep any analysis and reporting on track, and make sure that you’re delivering value through actionable insights to your clients.
For me, I just bought a new iPhone, so I think I’ll be staying away from stairs for a while.

With more data available than ever, how confident are you that your data analysis and recommendations are up to scratch? In this post I cover 5 simple checks that will help you deliver valuable, data driven insights to your clients.
Not enough of an intake for your Big Data macros? Find out more about Data Management here. Follow us on LinkedIn today!