Licensing SaaS (software-as-a-service) products can be complicated. Every vendor wants to encourage commitment. Commitment comes in long-term contracts of one or more years, with better prices associated with longer commitments.
Clients ask InterWorks to help them make contract capacity and term commitment decisions. Once we have enough usage data, it is easier to model with certainty. System usage during the few sprints (2-3 months) will enable the collection of comprehensive usage data. Work during this time will generate more system usage than usual due to testing new extract, load and transform (ELT) processes, building the data warehouse and developing your initial dashboard designs.
But, development activity will provide insight into normalized usage levels. This activity will provide enough usage to inform contract negotiations with your cloud platform and software vendors.
After you’ve completed a proof-of-concept (POC) on a significant number of your data workflows, you will begin to appreciate how to manage these costs. Five items will shape contract prices:
- Data loading frequency, size and speed
- Data transformation complexity and speed
- Database architecture testing
- Dashboards development and testing
- Cloud platform contract options
You will have to negotiate at least four contracts. Selecting a single software and platform vendor for your needs is not recommended for reasons I’ll discuss later. SaaS licensing reduces initial costs but requires significant contract commitments when considering two-to-five-year time frames.
The Single Cloud Vendor Approach
While it may sound appealing to go with one of the three prominent cloud platform vendors (Amazon, Microsoft and Google) for your entire software stack, this approach comes with three significant downsides:
- Single vendor lock-in
- Suboptimal software components
- Ongoing cost issues
For many years, legacy database vendors grew, at least partially, by acquiring startup technologies and expanding their offerings vertically and horizontally. I call this the buy your way to greatness model, which unfortunately doesn’t work well for customers. Why?
Over time even the best-laid acquisition plans produce less-than-expected synergies. The leaders and most-talented staff of the acquired companies leave when their non-compete clauses expire, and the formerly great company becomes mediocre. Big-company bureaucracy ensues, and the whole enterprise devolves into less than the sum of its parts.
Hoped-for synergies and improvements are only partially realized. Innovation slows. Sometimes entire product lines are discontinued to favor something new and presumably better. I’ve seen this happen over the past 25 years with many well-known technology brands.
Prominent cloud vendors are not immune to this phenomenon. They all want to find ways to complete their cloud offering so that you believe it will be easier to contract with their cloud platform and SaaS software. This approach is inherently suboptimal. Buying all your software from a single vendor will only give you access to some (but not all) of the best tools for your particular needs.
Single-vendor stack solutions typically result in performance, quality or cost trade-offs that may not be desirable to achieve contract simplicity.
Because the web is standards-driven, with many new vendors vying for a piece of your data workflow, going with a best-of-breed approach is better because you can:
- Select the best tools for your particular needs.
- Obtain better leverage in contract negotiations.
- Enjoy lower costs.
Selecting different data loading and transformation tools for different workflows is advantageous. You will likely have a primary enterprise extract, load and transform (ELT) tool that will handle most of your needs effectively. You may need a few other ELT tools for special situations. Selecting the best database for your workloads and billing preferences is essential for performance considerations and cost control (how are usage and storage capacity billed). Picking the right dashboard software and platform will save thousands of hours. The trade-offs are always between software costs and your labor time. Err on the side of saving time.
Selecting best-of-breed solutions may add software expense. It will also save thousands of hours of ongoing development costs that will compound over the years. The duration of each contract can range from zero commitment (higher unit cost) or longer time commitment (lower unit cost). Only commit to a single vendor when testing convinces your team that the vendor will meet your current and future needs. Then challenge your consulting partner to talk you out of the arrangement.
After selecting the cloud platform and software, the best way to determine the best mix of contract term length and unit costs is to sign a short-term contract and accept higher prices for the first 4 to 12 months of your project. Knowing your usage patterns provides the best foundation for negotiating longer-term agreements with your vendors. From a cost and performance perspective, this is the most prudent way to minimize long-term software costs and maximize the labor efficiency of your BI system.
In the next post, I’ll discuss how to determine what sprints should follow the successful completion of your proof-of-concept sprint.