Every marketer today recognises the value of data exhaust from their variety of tools, and yet they all struggle to make sense of its rising tide. This capability gap has fuelled the rise of a new profession, which the Harvard Business Review has called “the sexiest job of the 21st century”-  the data scientist.

If “data is the new oil” as industry analysts quip, then these data scientists are its petrochemical engineers, capable of refining and extracting value from the raw grist of ad impression logs, web traffic statistics, and household purchase data. Yet data science need not be the province of data scientists alone. It is often said that the most valuable trait of a good data scientist isn’t technical skill, it’s curiosity. The same might be said for the best marketers, and who is better equipped to pose questions about what strategies are successful than marketers themselves? 

A new breed of software is emerging that enables marketers to explore their data with the same facility as a Ph.D. programmer, leveraging something we’re all familiar with: a web-based user interface. Self-service data science, enabling a point-and-click approach to data exploration, visualisation, and analysis, is gaining ground for several reasons:

1. Direct access to data means faster time to insights

When a marketer has to channel data reporting requests through a data science team, it can take days for that request to be fulfilled.  Data bureaucracy is eliminated with self-service tools that put data at the fingertips of users.

2. Easy access to data means wider adoption

When data is hard to get, or requires a knowledge of programming language (like SQL) to query, it informs fewer decisions.  Conversely, when data is readily available, it becomes a natural part of more decision paths.

3. Web-based tools promote collaboration

When data is hidden in databases that only high priests can access, and a static data set is provided as an Excel file, collaboration can be difficult.  Web-based tools mean a URL about a successful campaign can be shared with a colleague, who in turn can take the analysis one step further. Collaboration and web-based software go hand-in-hand.

4. Campaigns never sleep, so the best data is fresh data

When data reports are pulled and shared via email, they – and the insights that are based on them – can quickly become out of date.  Web-based interfaces offer the promise of being connected to live, continuously updated data – and the ability to set alerts with triggers.  As marketing continues to go real-time, knowing within minutes that a digital campaign is in a nosedive translates into meaningful dollars.

Build, partner, or buy?

The power of data analysis comes when multiple data sets can be brought together- when media spend can be connected to CRM files and in-store purchases. Rarely are these data sets to be found from the same vendor. The software that marketers rely on for data science should thus be distinct and separate from software for media execution.

The question then becomes: Do we build these self-service tools ourselves, or should we buy them? Depending on the scale of data you’re working with, whether you already have a data warehouse or are looking for an end-to-end solution, and the level of customisation needed, there are a number of vendors to choose from: Tableau, SiSense, ZoomData, Looker, Spotfire, to name just a few. (At Metamarkets, we specialise in working with high-dimensional, large-scale data sets, typical of programmatic media buying.)

As a former CTO that now works regularly with marketers, my overwhelming advice is: You may not choose my solution, but don’t attempt to build this entirely yourself. Fast, flexible user interfaces that enable visualisation and analysis of data can be deceptively difficult to build. And while these tools are sexy to create, and thus make for appealing projects for internal engineering and product teams, they are rarely a company’s core competency.

It’s a far better choice to investigate existing vendors, consult with colleagues who have successfully deployed a self-service solution, and get something running within a few months.  The alternative is to devote several quarters to a grand vision that is unlikely to succeed.

The rising tide of data, which began a few years ago, is only going to increase as marketing goes ever more digital and television becomes addressable. Marketers that have intelligently navigated the data challenges of today are likely to find themselves as the CMOs of tomorrow.