A common denominator for pricing and negotiating Business Intelligence (BI) and Analytics software

BI and analytics software packaging and pricing are a Wild West with few common practices among the vendors. Comparing and contrasting vendor prices and negotiating with vendors is challenging because

  • Few vendors publish list prices, so when a vendor tells you you are getting a certain discount you can’t really verify whether the discount numbers are valid or not.
  • Vendors base their prices on multiple variables such as
    • Total number of users
    • Concurrent users
    • User types
    • Connectivity to certain types of data sources
    • Number of CPU cores or sockets
    • CPU clock speed
    • Amount of RAM
    • Server Operating System (OS)
    • Environments such as development, test, QA (quality assurance), UAT (user acceptance testing), production, and DR (disaster recovery)

So how do you know if you are getting a good deal? Here’s a best practice and a few price ranges you can use to get you started. First of all, at the end of the day, it’s the number of users and user types that are always a common denominator regardless of BI software platform or your  particular implementation. Consider the following price ranges for specific user types

  • Static report viewers - PDFs, etc - $0-$50 per seat
  • Interactive report viewing (sorting, filtering) $500-$1,000 per seat
  • Advanced interactive report viewing (pivoting, ranking, slicing/dicing) $1,000-$2,000 per seat
  • Power users, report authoring $2,000-$2,500 per seat
  • Data scientists, predictive/advanced analytics users $5,000-$10,000 per seat
  • Developers/Administrators $2,500-$5,000 per seat

Next, do some math to figure out a ball park number that fits your BI project particulars. Then start your negotiations based on the following discounts

  • 10%-50% discount if your deal is in low 6 digits
  • 50-70% discount if your deal is in high 6 digits
  • 70+ discount if your deal is in 7 digits

Last, but not least, IMHO, BI vendors should not charge for more than a single environment, so dev/test/QA/DR should be included in all BI deals since that's part and parcel of how all enterprise software is deployed and used.

Looking forward to everyone's comments if the ranges you see here relate to your field experience with BI software prices


Translate to SaaS?

Hi Boris,

Thanks for this. One thought that came to mind was that while I was a bit of a critic on (lack of) published pricing in the past, given the amount of integration work you and others discuss, and I see in the field, when I look at some of the published prices frankly they are suspect as we have some idea on the overall cost which is in some cases I expect are many x that of product price, perhaps 10x in some cases.

I may comment further later but just had a quick question--wondering if you could clarify whether/which are one-time events, annual, etc. and also anything you care to share on how this is generally translated to annualized SaaS subscriptions or cloud rental. Nothing specific, just fishing for general view from your perspective, broad trends, etc. --anything you can share.

Appreciate this post. - Thanks, MM

Mark, great questions.

Mark, great questions. Obviously one has to add anywhere between 18% and 28% in the annual maintenance costs that BI vendors charge. I don't think I've seen 10x multipliers for BI services (unless we go really broad with EDW, ETL, DQ, MDM, etc implementations). Just to deploy BI platforms I usually use 0.5x - 3x multiplier. Last but not least I am hoping my colleague Martha Bennett will chime in on mobile and SaaS BI pricing - she's currently doing deep dives into these market areas. Cheers!

Thanks Boris--additional thoughts

Boris, yes I was thinking broad system installs at 10x -- ERP that includes BI, ETL, MDM, etc. As you know Kyield provides enterprise-wide data management, BI and analytics, with elements of MDM, discovery, and productivity, so while we don't supply major components, we do provide the brains that makes sense of the enterprise and so must integrate while pushing towards a more interoperable approach with adaptive data management.

For example--this may have some value to this topic, we have priced on a pp annualized basis for very large enterprises, with significant incentive to scale to entire org as it's proven very wise to do so for the biggest ROI areas, but only after forecasting integration costs, and with a solid / conservative growth plan starting with one or more business units of sufficient scale. Now in the main case I am referring to we were not providing the SI, which priced separately by preferred customer SI, but since the Kyield scenario provides value for the entire tech stack--and also causes project costs, it's very important to me for expectations and cost estimates to be accurate.

It has become a bit easier over the last couple of years to price such projects, in part as experience grows and in part due to increased support, maturing underlying stack, and key additional options (your Future of BI report was published after all just over 2 years ago if memory serves...).

That said, I am still a bit concerned over the carrot approach to published pricing in related areas given that 90% of data analytics or so is still in services (according to major published analyst reports anyway), and only a small percentage automated, with SW licenses & 'shelf' pricing really quite the minority. Obviously, business models that depend on this carrot approach are one of the reasons so many projects are either killed too early to provide value, or struggle to provide a durable & consistent ROI in eyes of many customers.

So while confident our direction and intent are aligned, I'm just a bit worried about exploitation of good intentions. One thing to price a product--quite another to price the costs & impact to entire org the product creates on an ongoing basis, and still quite another to capture accurate ROI.

No question it's an important topic, and somewhat fluid. - Thanks again, MM

SaaS Licensing Considerations

I think you'll find with SaaS vendors, that we try to accommodate most of Boris' recommendations, but spread costs more evenly across a three or four year timeframe, and make it more closely match what you are actually consuming (or going to consume). Some vendors charge you an hourly rate like AWS does, which still leaves most of the deployment complexities-like integrations and data loads, in your hands.

The difficulty of course, is that we don't know what the complexity of your deployment is going to be until we begin managing it (so you don't have to). As Boris points out, we do want you to consider all of your costs across the project including ETL, EDW, Quality, labor, etc. So we will ask questions like how much data are you planning on analyzing, or how much will we need to store? How frequent and complicated are the transforms, and will they be fed by accessing cloud-based APIs, on-premise JDBC or data uploads, or complex Map Reduce processing? And finally, similar to on-premise vendors, who are the consumers? Are they employees or external customers? Viewers or Ad-hoc editors?

Once we know this kind information, it is much easier for us to show you not only what is the current value that you are negotiating--based on size, complexity and consumption, but also show you how that value will scale over time as your deployments grow.