I’ve been lucky enough to be at Fivetran over the past two and a half years as we’ve grown our company size fivefold and increased our valuation tenfold. We've moved from Series A through Series C funding and officially crossed the border into unicorn status. Along the way, through multiple positions in sales and then marketing, working across different teams, I’ve noticed one key to our success: the use of data to drive decisions.
I used our data tools when I was in sales, but it wasn’t until I joined marketing that I was able to fully appreciate the power of comprehensive analytics. I got further into our modern data stack and soon realized how deeply we could understand marketing performance with all the relevant data consolidated in one place, ready to access for ad hoc insights.
Bridging the Gap Between Product, Marketing and Sales Analytics
It’s fairly easy to use the interfaces of our various marketing data sources to generate top-of-funnel metrics such as clicks, costs and conversions. Source UIs will quickly get you to a very general understanding of how marketing campaigns are doing. In fact, I tried to take this obvious path on my first project after joining the marketing department.
The scope of the project involved a typical email campaign to encourage our customers to try a particular product feature. Initially, I only identified Marketo interaction metrics, such as opens, clicks and unsubscribes.
During a review of the results, my colleague and trusted advisor Joseph Long, Senior Digital Growth Manager at Fivetran, gave me some honest feedback: Delivering this as finalized project results would be similar to delivering “fluff metrics.” That is, we could get a high-level understanding of how well the marketing campaign did, but we wouldn’t know whether we had influenced revenue by driving a specific set of intended actions. In other words, we wouldn’t be able to determine whether we’d accomplished our goals.
Luckily, outside of the default reporting found in the Marketo UI, we had already been centralizing all of the relevant sources into our data warehouse, where data was continually being refreshed by our Fivetran instance. In addition to our production database (PostgreSQL) hosting product events, including feature and connector usage, we also used:
- Google Analytics to track our website traffic
- Pendo to track product events as well as display in-app notifications
- Salesforce as our primary source of truth for sales
- Multiple other sources, including Google Ads, LinkedIn Ads, Facebook Ads, Google Sheets, Zendesk, webhooks and many more
This data made it possible to bridge the gap between marketing, sales and product analytics. Having so much data readily available to join helped me answer questions such as:
- What were the most efficient ads in terms of influencing targeted user actions?
- What were the most efficient email campaigns in terms of customer outreach indicated by additional product feature usage?
- What were all customer touchpoints at the top-of-funnel across all pertinent channels?
- What types (segments) of customers should we target for which feature sets, using the expected business value returned by their product usage?
And so many more questions that we’re continuing to iterate on! Lighthearted warning here: As your data analytics team continues to succeed, and you create internal champions of a data-driven approach, more departments will take note of this success and the volume of requests to your analytics team will naturally rise. Here’s how Terence Knorr, Senior Managing Editor of BI at Braze, describes this transition after adopting Fivetran and a modern data stack:
At our monthly town halls, analytics and BI are consulted or referenced consistently and measurable KPIs are becoming more critical. It has completely started to change the culture here and we have a line of people wanting us to do more for them. It’s a lot but it is extremely exciting.
Marketing Insights in Four Easy Steps
To readily access information across your marketing, product and sales tools on the fly, here’s a high-level overview of the steps required:
1. Gather a list of data sources — and understand how each source is used. Include the data that’s most relevant to your current and upcoming questions. At Fivetran, all of our relevant data sources are already centralized in our data warehouse and data updates at near real time frequencies — meaning that I could trust this data to be up to date for agile decision making. In the cases where a connector didn’t already exist in our production account, such as when we first implemented Pendo mid-last-year, I was able to set up an integration within 10 minutes.
2. Determine which metrics and KPIs are most relevant to your decision-making. In my case, I wanted to understand the ties between marketing sources (including email campaigns and paid social ads) to product actions (feature usage and downstream revenue implications, hosted in our CRM). When in doubt, think about how this ties back to revenue.
3. Understand the data model. You’ll need to understand what the data model looks like and how tables can be joined together. Prebuilt entity relationship diagrams that Fivetran had available for its data sources helped enormously. A Marketo example is below:
4. Test with an ad hoc query. Save time by implementing your theory into a dashboard. Whether you’re starting off with Google Data Studio or you’re a seasoned user of data visualization tools like Tableau, Looker, Sigma or PowerBI, you can often save your end users multiples of time by creating ongoing monitoring dashboards.
How to Start Modernizing Your Marketing Stack
Here are some options for modernizing your marketing analytics setup, whether you’re currently a Fivetran customer or looking to automate your existing data stack:
- For a fresh perspective on your marketing analytics program, feel free to book time with me through Calendly, or email me any questions at email@example.com.
- Sign up for a Fivetran trial: Every new account comes with 14 days of unlimited features, and the clock doesn’t start ticking until you’ve completed your initial historical sync. Many of our technology partners offer free trials as well, so you can easily test a complete modern data stack with seamless integrations.
- Try a new connector: If you’re already a current Fivetran customer, every new connector you set up comes with 14 days of free usage. You can freely test new sources and validate how you’ll use them in your reports prior to paying a cent.
- Reach out to your customer success manager (CSM): Current customers already have a dedicated CSM. Reach out to them directly via email to learn more about best practices, or file a support ticket to get in touch.
- Read how Fivetran customer Honeycomb quickly integrated finance, sales, marketing and product data to drive better business decisions.