4 Lessons from Meta: How to Break Down Silos and Improve the Efficiency of Your Analytics Organization
At Neustar’s Brave New Worlds data strategy and analytics event in November 2021, Neustar VP Marc Vermut interviewed A. Charles Thomas, VP of Data Science at Meta, to talk about what constitutes an efficient analytics organization. That session covered a lot of ground, and we invite you to check it out in full right here.
Their discussion on the topic of corporate silos—and the cultural shifts required to break them—was particularly interesting, especially in light of the accelerating deprecation of digital identifiers like cookies and mobile IDs. Many companies are newly investing in their first-party data assets as such, and in doing so, they’re creating a common language across their organization. In short, if marketing, sales, distribution, IT, and finance have a shared view of the customer, they stand a better chance of breaking down departmental silos and working more effectively together.
Read on to learn about Thomas' key views on what it takes to build a successful data science culture that can help break down the departmental silos that hamper growth.
1. "The biggest hindrance of all progress is people."
That’s as blunt as it gets. The biggest hindrance is not technology, processes, the economy, or a new competitor breaking in, but people "latching on to pet projects," as Thomas puts it, and refusing to see that the world has changed around them. Organizational silos encourage people to defend their own turf at the expense of company goals. And the incentive structure can be destructive if it's left unchanged. Allocating even 10% of someone's responsibilities and compensation to cross-functional projects can do wonders to spark innovation and cross-team collaboration.
2. Hire people with soft skills to build bridges.
Silos are full of very smart people who are ultra-specialized. Technical acumen is great, but Thomas reminds us that "analytical skills can always be farmed out." It's the softer skills that create true transformation. You want team members who can contribute, listen, and learn from one another. At Meta, the Global Marketing Data Science team has 15 PhDs, and they come from fields as diverse as behavioral sciences and social sciences. Thomas values people who can see beyond the data series, think holistically about a business problem, and thrive in a collaborative environment as a cross-functional team.
3. Silos are slow. Time to speed things up.
Modern marketing is complex. Data comes from all directions, and it needs to be cleaned up, onboarded, and activated across a multitude of marketing applications. Meta is running a sophisticated marketing mix modeling (MMM) optimization practice, with multi-touch attribution (MTA) and other models to "get more juice from the squeeze," as Thomas puts it. At any other company, that level of complexity could easily mean costly delays. But Thomas made fast deployments and short feedback loops a priority. That focus on speed forced different teams to embrace unified identity, and develop a data and technology infrastructure that could support that vision.
4. Align your marketing strategy to business outcomes.
Your job as a marketer is not to push a campaign out, but to deliver results that are good for the business and improve the customer experience. There are 3-5 key business priorities at Meta at any point in time, Thomas relates, and if the projects in the marketing pipeline don’t line up with those priorities, they get the boot. Take any company in the world, and it’s unlikely that the decision-makers at the top are fully versed in the math that underpins your new model, or in advanced analytics in general. Thomas advises that marketers need to build a story the C-suite can relate to, and make it crystal clear that their marketing KPIs are aligned with the business outcomes they care most about. When everyone is pulling in the same direction, corporate silos become irrelevant. Outcome-based marketing is a great place to start.