I just spent 3 whirlwind days at Google Cloud Next ‘19, Google Cloud Platform’s (GCP) annual global conference, held in sunny San Francisco. The sheer volume of information was mind-boggling and I’ve barely had 48 hours to digest it all. But looking back, three clear themes emerged that excites us at Data Runs Deep.
- Lasering in on Marketing Analytics
- Democratisation of Data Science
- Easing the ingestion of data into GCP
Theme 1: Lasering in on Marketing Analytics
Data Runs Deep has always had marketing analytics at its core. It was exciting, therefore, to see that Marketing Analytics and Data Warehouse Modernisation are two key areas that the Google Cloud team are prioritising this year. This means lots of exciting features (which I’ll get to in a minute) that will allow marketers to be more effective in their jobs.
The upcoming Cloud For Marketing console will make marketing data science more accessible to marketers. Cloud For Marketing includes features that make it easy to ingest data into BigQuery, pre-prepared SQL statements (no more trying to figure out UNNEST statements!) for customer and revenue analysis, pre-prepared predictive models, and Google Data Studio templates for reporting. It is, in essence, a hub for all of your marketing analytics needs. The aim is to enable marketers to understand their customers better by breaking down siloes to consider the whole picture, and to enable richer data-driven segmentation.
Tied to this richer marketing analytics is the underlying modern data warehouse. Google BigQuery keeps going from strength to strength. First, it’s becoming faster: a 1 petabyte query took 4 minutes to run in 2016 and, just the other day, the same query ran in 11.7 seconds! How long will it take next year? Second, the focus is now on real-time analytics on streaming data. In the future we won’t be thinking of data as something that “happens” in batches, but as something that is fluid and constant. BigQuery is readying itself for that future.
Speaking of speed and real-time, BigQuery BI Engine is an in-memory analysis layer over BigQuery that provides sub-second responses! This means that your queries will run super fast and stop complaints from impatient analysts. Currently BI Engine can only be accessed via Google Data Studio; but I’m sure it will move further afield soon.
Theme 2: Democratisation Of Data Science
A key theme in Day 2’s keynote was about bringing artificial intelligence (AI) to everyone. AI should be accessible to not just data scientists, but also to data engineers, business stakeholders, and decision makers. While we still have a way to go before data science (or AI or ML) is ubiquitous, Google Cloud is paving the way.
AutoML Tables allows you to build predictive models without having to write a single line of code or SQL! Simply point AutoML Tables to a Google BigQuery dataset, select what column you are predicting for, and away you go. Predictive modelling has never been easier!
If AutoML Tables is too easy for you, and building a custom TensorFlow model is way too hard, you might want to consider BigQuery Machine Learning (BQML). BQML brings machine learning directly into Google BigQuery. It abstracts away the painful parts of machine learning (e.g. feature engineering and hyperparameter tuning) to allow non-machine learning specialists to get to a working model sooner. The upcoming features that interest me are the introduction of new types of models and the ability to import TensorFlow models.
While on the topic of BigQuery, the Google Sheets data connector for Bigquery will soon be generally available. It allows Sheets users the ability to easily query and manipulate BigQuery data within a familiar interface. You can also apply Sheets formulas to the imported data. No SQL skills required!
Finally, there is the AI Hub [video], Google’s one-stop shop to discover AI assets. You can find pre-prepared TensorFlow modules, notebooks, trained models, and more. This means we can get started on AI without having to start from scratch! You can also publish assets and privately share them with your team for greater organisational efficiency.
Theme 3: Easing The Ingestion Of Data Into GCP
Everything I have discussed so far relies on the presence of high-quality data that is accessible by GCP’s data and AI services. Getting data in, however, tends to be that first hard step that can stall progress. Thankfully Google announced a number of new services that now makes it almost ridiculously easy to get your marketing data into GCP.
BigQuery Data Transfers has now extended from just Google products (e.g. Google Ads, Campaign Manager, etc.) to supporting 136 third-party data sources! This means it is now even easier to bring your marketing data from Salesforce, Facebook, Twitter, Instagram, Criteo, Adobe, etc. into BigQuery. From there you can run a holistic analysis on your marketing campaigns, visualise all marketing channels in Google Data Studio, and more!
If you need to transform your data (e.g. remove PII, perform custom calculations on a combination of fields, etc.) before storing it in BigQuery, then there is Cloud Data Fusion [video]! Cloud Data Fusion allows you to connect to several data sources, including on-premise databases and popular CRMs, configure your data transformations, and then store them in BigQuery (among several other options). So what? I forgot to mention that you can do all of this without writing a single line of code! Easy!
(Bonus) Theme 4: What Is The Future Of Our Jobs?
I was asked the same question several times with regard to the above announcements: “Do you think data analysts’ and marketing analysts’ jobs are at stake?” My answer was a resounding “NO!”.
If anything, I believe this is the start of the next evolution cycle of the digital/data/marketing analyst’s role. By taking away the pain of bringing data into the cloud, abstracting away the complex features of machine learning, and not having to start every AI project from scratch, we are poised to use 90% of our time and 90% of our brainpower on solving our business problems. Solving these problems in quicktime will allow us to turn our attention to more interesting and richly complex problems; which in turn leads to much richer career development paths. Exciting times are most definitely ahead!
Here are some useful resources that will help you get up to speed with all the announcements from Next
What Next After Next?
It’s safe to say that at Data Runs Deep, we’re extremely excited at how the digital marketing landscape is going to change over the next year thanks to the Google Cloud Platform. The fact that we can answer marketing questions sooner and, more importantly, solve marketing problems faster is a mouth-watering one. As a consultancy, Data Runs Deep is already well-poised at the intersection of the Google Marketing Platform, Google Analytics 360, and Google Cloud Platform - and the team are giddy at the prospect of being on this ride and seeing where it takes us and our customers.
Keep an eye on our blog and The Flying Beagle for more in-depth write-ups of the above products over the coming weeks.
If you would like to find out more on the above, please say g’day and we’re always happy to have a chat.