Industry pundits say that 85% of big data projects are failing, and that 84% of digital transformation projects are failing. Understandably, there are a lot of CIOs and CDOs out there who are getting anxious.

The reality is that big data and digital transformation projects fail for the same reason that other IT projects fail: there isn’t a clear business case, users can’t seem to visualize the project and understand how it will work, requirements aren’t clearly defined, and project deadlines are missed.

Vendors with a “hit and run” policy of selling their products to a client and then leaving the client to figure out how to work with the product, deserve part of the blame.

SEE: Managing vendor relationships: Time commitment, benefits, and pain points (Tech Pro Research)

This is what got me excited when I spoke with SAP last week about the company’s new Leonardo approach to selling big data, IoT, digital transformation, and analytics products.

Named after the renaissance artist, designer, and innovator Leonardo da Vinci, Leonardo is a sales methodology that aims to collaborate with a would-be or present client to address opportunities and business problems that are worth solving, and then come up with the right set of solutions that can rapidly deliver a demonstrable business result.

“A primary goal of the process is to place ourselves in the customer’s shoes, and to work closely with the customer,” said Jorge Granada, global lead of Leonardo marketing at SAP.

Granada said that by using the Leonardo methodology, SAP could sit with a customer or prospect and:

  • Perform discovery and identify business opportunities or problems to solve in one day.
  • Design a preliminary prototype of a project in as soon as one to two weeks.
  • Produce a functional prototype for a project in six to eight weeks which can then be fully developed into a deployable solution.

The process is expedited by using industry accelerators, which SAP describes as pre-vetted solution sets that already are tailored to the needs of specific industry verticals.

“What we want to do is to help the customer develop a blueprint for action and also visualize the types of big data, IoT, analytics and other solutions that will benefit the business, and how they will benefit the business,” said Granada. “If we don’t feel we can help the customer, we will tell him that, too.”

Hopefully, other vendors will follow suit.

Meanwhile, what should IT and business decision makers with big data, IoT, and digital transformation goals be doing?

1. Avoid hit and run vendors

If your vendor doesn’t offer collaboration that helps you define, refine, develop, test and implement your project goals, look for one that does. This means avoiding vendors that sell you a solution and then disappear or become unresponsive when you need help or support.

SEE: How to choose and manage great tech partners (ZDNet special feature) | Download the PDF report (TechRepublic)

2. Start small with IoT

Internet of things technology is still in early stages of adoption. If your company lacks internal IoT expertise (and many companies do), it’s best to start modestly with IoT, build upon some successes that directly contribute to the business, and then expand. IoT is an area where you want to have strong partnerships with the vendors providing tools and solutions.

3. Practice zero tolerance for project failures

Businesses are rapidly approaching a tipping point where CEOs and boards aren’t going to tolerate any more digital transformation or big data project failures–and there’s no reason why they should. There are enough methodologies and best practices out there to ensure that projects are well-run, and that plugs are pulled early if a project isn’t going to work. Most importantly, what you do with a big data (or any) project must resonate with the business. IT’ers, and those without an IT background, must be able to visualize and then see the result that the project delivers to the business. If your project can’t pass these fundamental litmus tests, it’s time to question why you are doing it.

Also see: