How can you avoid wasting engineering cycles on data pipeline creation and maintenance? Powered by Fivetran.
As a product manager, I’m constantly looking for the magical low-investment, high-payoff feature or enhancement. In other words, I want to know how I can improve the success of my product without wasting engineering cycles.
In my last product role, I was building data-driven tools for musicians, trying to crack the secret to advertising, music and social media and create a predictable recipe for musical stardom. The basis of the hype machine was data — pulling data from streaming services and ad or social platforms. Our team invested months in building custom ETL to get the data into our product before investing months more trying to create actionable insights for our users to grow their musical brands.
Looking back now, with my time as a product manager at Fivetran under my belt, this investment of countless weeks of time to get data from common sources like Facebook Ads, Facebook Pages, Snapchat Ads and Twitter Ads feels like a waste. It significantly increased time to value for our users — and that time was a critical cost for a small startup with big aspirations. Cracking the hype machine code is no joke.
Getting the data was just step one. The real work came in the analytic modeling and building the related user experience. At the time, that was the obstacle that me and my engineering partners saw in front of us — a required step to create the big payoff. Now I know that Fivetran, with our REST API and Connect Card offering, can help engineers and product managers skip these cycles and move straight to the good stuff — creating value for users.
I’ve worked on several build-vs.-buy analyses in my tenure as a product manager, and I’ve always followed a simple three-step process:
1. What are my must-have requirements in this functionality?
Boil it down to the goal of the project and the data and functionality you’ll need to support it. Things to consider for data integrations:
Will I get all the data I need to accomplish meaningful analytical insights for my customers?
Will I get the update frequency I need to delight my customers?
If something breaks, how quickly will it be addressed?
2. How much will it cost to build vs. buy?
Get quotes from several vendors and determine how that cost will scale over time based on usage projections.
Get estimates from engineering for building the integrations, then double it and factor in maintenance.
Get estimates from engineering for integrating with the partner, then double it.
Opportunity cost: Am I giving up key differentiators or time to market by building this with my team?
3. Can you trust the vendor?
Understand the company’s funding: Do they have longevity?
Have you been able to work with their product and support team and felt a sense of partnership?
Do you trust they’ll grow their product to match your future needs? A related question: Do they have the engineering resources to do this?
I use this process to make important product decisions, and I hope it will help you determine what works best for you and your product, particularly when it comes to SaaS data integration.
Now go forth and conquer the ever-expanding market of data solutions!