Kevin Smith has led embedded analytics teams and designed SaaS products for companies such as Birst, IAS, ServiceSource, and SAP Labs. Putting that experience to good use, he is now principal consultant and founder of NextWave Business Intelligence. A Colorado-based analytics product strategy firm, NextWave helps SaaS companies incorporate business intelligence into their software products. Sometimes this means building a BI solution from scratch, and other times it involves surveying existing solutions for the right fit.
We got in touch with Kevin to learn more about his methodology for navigating the increasingly complex business intelligence market.
Q: What would you say are the top three mistakes companies commonly make in purchasing embedded BI?
The mistake I see most often is that, when purchasing analytics for embedding—specifically for creating a customer-facing data product—companies fail to consider the differences between enterprise analytics and embedded analytics use cases. For example, purchasers might get distracted by beautiful charts and fancy user interfaces (both of which are important) but forget that they will need to manage the embedded analytics over the product life cycle. Is it easy to add new customer tenants? Can the analytics be white labeled? Can customers create analytics on their own or do they need assistance for each change? All are essential to embedded data products but might not show up on a decision matrix tailored exclusively to enterprise analytics.
Purchasers also fail to consider the need to support personas within their analytics. You won’t have a generic “user” visiting your data product. Your user will have specific objectives, work flows, and needs they are trying to address. Creating a “one size fits all” analytic solution is a recipe for failure. You must anticipate the creation (and support of) multiple personas and ensure that the platform you purchase can support this need.
Lastly, pricing. To create successful data products with embedded analytics, it’s essential that you generate widespread adoption and engagement. Buying a platform that charges on a per-user basis naturally discourages the creation of free or basic analytics tiers—the gateway to high-margin analytic plans. Make sure that the platform you choose has pricing that allows you to offer analytics to everyone without breaking the bank.
Q: What red flags should companies look for before investing time in evaluating a BI vendor?
For me, a big red flag is when a vendor claims to offer embedded analytics for data products but doesn’t really understand what is required to build, deploy, and maintain customer-facing analytics. When a vendor enables simple iframe embedding but lacks the ability to embed specific analytic elements alongside core application content, it’s a sign they don’t understand the needs of a product team. If the vendor doesn’t have tools that allow for the management of the “product ecosystem,” including on-boarding customers, deployment, rollback, and operation of customer instances at scale, same thing. It’s a big red flag indicating a mismatch between the platform functionality and needs of a data product team.
Q: How can companies identify BI vendors who take a partnership approach to the business relationship?
The best way to explain this is with the story of my first data product. Several years ago, I narrowed the list of suitable analytic vendors for a product down to two candidates—neither of which I’d ever used before.
One of those vendors did everything possible to make the process easy and fast, standing up proof-of-concept instances that we could use for evaluation purposes and helping us to think through considerations that might not have occurred to us otherwise.
The other vendor? We never got through the proof-of-concept phase because, rather than making the process easy, they put every possible obstacle in our path. They were dismissive of our business model, ignorant of our business needs, and demanded that we “read the technical documentation” on their website rather than answering our questions.
Guess which vendor we chose?
I look for vendors that are eager to help meet a business need rather than to simply close a deal before the quarter ends. Setting up meetings to discuss the business model and offering suggestions both about the plan and how the technology might fit is a great start. Giving examples of how other customers architected their data products—including things beyond the technical like pricing and go-to-market strategy—show that the vendor is thinking about more than just a sale.