This confluence of legacies has created a barrier to the innovation needed in a world of softwarization. Obviously, change is needed. Change is never easy, but if we can develop a consensus on a path that overcomes these barriers, it will be easier. 

The Shape of the Solution

Start-ups and small companies have demonstrated an ability to provide the innovation stream that is so desperately needed. They are often encouraged to demonstrate in unpaid PoCs. These PoCs can be important in providing both the innovative team and the CSP an opportunity to learn how to work together.  But, for the reasons described above, such PoCs are generally not in themselves paid engagements and do not lead to paid engagements. Payment is reserved for the RFP process.   

A solution then, involves paid on-ramps for innovation.  One way these on-ramps can be created is through a series of funded steps.  Such funded steps might include:

1. Funding early technical and business case studies
2. Proof of concept demonstrations
3. Lab tests
4. Paid field tests
5. Small deployments

In this context, it must be clear that not all efforts at each stage will be successful.  This means that CSP managers have to be evaluated not on the success or failure of a particular effort, but rather on:

• Degree of innovation sought
• Professional handling including clear and complete documentation of requirements; development and adherence to a realistic schedule; clear, measurable criteria for success; documented test plan for measuring success; and definition of next step if success is achieved.

In this, there needs to be some sense of portfolio management. In many VCs, a 90 percent failure rate is expected. The 10 percent of successes pay for everything else and create above-average returns on investment. CSPs may expect more than a 10 percent success rate at every one of the five steps above. But even if there is a 75 percent success rate at each of the five steps, the conjoint probability of a particular effort starting at step 1 and achieving success at step 5 may be fairly low. Management has to understand, accept, and support this throughout the process.

The next question is how to fully onboard the proven innovative technology. Here there are two possible approaches. The first requires the CSP to break its acquisition process down to smaller and different types of acquisitions. For example, instead of buying a “complete” xG infrastructure, the CSP may need to acquire separate subsystems in a series of separate acquisitions. RFPs may be appropriate for these.  But it may also be clear a sole source contract makes the most sense based on a successful end of the five steps above. Also, most innovative software today is sold as Software-as-a-Service (SaaS). This looks to a CSP like an expense item and not as a capital expenditure, whereas infrastructure is generally considered a capital expense. As a result, some accounting adjustments may be needed

A second possible approach is for the CSP to work with one or more of the large vendors to get the large vendor(s) to acquire the innovative technology to deliver to and support the large CSP. This requires the least change by the CSP. It also provides a sustainable path for the large vendor. But it can run into the not-invented-here syndrome in the large vendor.  So, while this appears to be an attractive alternative, it is fraught with difficulty.

Fostering Innovation

Of course the devil is in the details. Let’s consider what’s important, though. By providing funded projects, the lifeblood of small innovative companies, CSPs can create evolutionary niches for successful start-ups and small companies. And in so doing, they create a sustainable innovation ecosystem that can provide the innovative software that they themselves so desperately need.