Harnessing the Data Cloud to Empower Our Own Marketing Team: Building a Digital Ads Ecosystem on Snowflake

You need metrics to do your job well as a marketer but getting clear, meaningful metrics is a huge challenge. While digital advertisers and paid media professionals are on the hook to build ample sales pipeline and maximize return on ad spend (ROAS), they’re also expected to deliver personalized advertising content while navigating evolving privacy requirements and adhering to consumer expectations—all while extracting insights from siloed ad platforms. 

According to Salesforce’s 8th annual State of Marketing Report, marketers cite measuring marketing ROI and attribution as their second biggest challenge—something that we here at Snowflake can certainly relate to. Ongoing changes in data privacy and policy have had a significant effect on marketing analytics and how our marketing team operates. From the EU’s General Data Protection Regulation (GDPR) to gearing up for a cookieless future, we’re all in for a rough ride. We’ve been reliant on the cookie to identify people—we haven’t even deprecated the cookie and we’re already seeing challenges. 

Combining insights from the ever-growing adtech ecosystem is crucial, but silos and privacy regulations make it hard to extract and work with all that data to create a complete, unified picture of the audience. Everyone in digital advertising has accumulated their own data. Due to privacy regulations, we can’t just share that data outright, so we have to find anonymized ways to extract insights. 

Personal identifiers such as name, email, and employer have to be removed. A digital ad partner can’t tell you, “This is Bob Smith, he works at Company X, with job title Y.” Instead, our target, Bob, is grouped anonymously into a cohort with his specific attributes. The ad partner shares that this profile works at a certain kind of company, like enterprises with over $100M in revenue, with certain skills (like Tableau, Streamlit, Python and Marketo integrations). This gives the marketer enough information to target meaningful campaigns without having to receive PII from ad partners. 

Here’s how Snowflake’s digital marketing team creates a full-funnel view of our ad performance in four steps: 

1. Ingest data from the 8 ad platforms with Fivetran and normalize the data sets on Snowflake.

2. Model data transformations with dbt to unify the data.

3. Provide normalized data sets through reports, dashboards and apps powered by Tableau, Snowsight and Streamlit.

4. Share refined and enriched data back to the ad platforms so they can fine-tune their algorithms.

A lot of the pre-processing done by Fivetran and use of dbt packages in those first two steps accelerates our time to deployment from weeks to one or two business days. Now our team can run an ad to profiles like Bob’s, measure click-through rates (CTR), and compare ad performance on each platform on an apples-to-apples basis. That means our Marketing Intelligence team can focus on integrating the ad data with our internal data to match it to campaign responses.

As far as cookies go, at Snowflake we had the foresight to capture the full URL of the campaign source—it captures whether the lead came from Facebook, paid search, the blog or another source. This means we can parse the URL to get insights ourselves. This collaborative effort between the Snowflake Marketing Intelligence, Marketing Operations and Digital Marketing teams lets us integrate URL-based insights with data from our entire ad ecosystem. 

Our data is governed in Snowflake, so we can share it with anyone across the company—contrast this with the small group of marketers that would usually have direct access to data stored in platforms such as Facebook or LinkedIn, and even fewer with access to all the ad platforms. Snowflake has made data accessibility a non-issue. Resources scale up to accommodate compute-heavy calculations or scale down for the average day, without requiring human intervention. The data is ready on-demand and the system just works.

It’s hard to derive insights from the onslaught of data, but it’s more important than ever. Marketers are under ever-increasing pressure to show results. HubSpot’s State of Marketing report finds that 92% of marketers expect their budget to stay the same or increase in 2023. So the perceived scrutiny on marketing budgets isn’t a budget-cutting conversation, it’s an impetus to be more data-driven. In this world of endless data, we’d just better be able to justify ourselves. 

Here at Snowflake, we are the first customer for our own innovations. We can predict real-time ROI to dynamically optimize marketing programs and ultimately improve upon B2B marketing analytics practices. For the Digital Marketing and Marketing Intelligence teams, our mission is to build scalable solutions and models to help Snowflake Marketing become the industry’s most insights-driven team. For more on how our Marketing Intelligence team stays on the cutting edge, join us for our webinar, Building a Robust Digital Ads Ecosystem to Maximize ROAS.