When trying to develop innovative technologies, engineers must be willing to take risks, make mistakes and move ahead incrementally without having all the data.

Jeff Finkelstein

February 17, 2015

7 Min Read
Can't Succeed Without Failing

Dying is easy. Comedy is hard — Attributed to Edmund Kean, 1787-1833

Dying is easy. Comedy is easy. Innovation is easy. Selling it to business partners is hard. — Me, 2015

Whether Edmund Kean (a Shakespearean actor born in London) actually said the first statement remains an open question. My paraphrasing of it is factual as it is now in print. But whether my quote has value remains to be seen.

I have spent the better part of my 20-plus years in cable working on innovative technologies. Starting with pre-DOCSIS LanCity, Com21, CDLP technologies, then onto DOCSIS 1.0 through to todays DOCSIS 3.1, I have been involved either in developing best practices or deploying, managing or writing specifications as part of the CableLabs spec teams. No one technology was easy to create or deploy, but with each successive generation we have had to start at square one each and every time -- pitching it, selling it, convincing the business side to let us build what we believed was necessary to grow our capacity to meet the ever-increasing demands of our customers.

I remember well taking the idea of running gigabit services over coax to CableLabs and a few vendors back in 2008. (It was initially called AMP for Advanced MAC-PHY.) My team had done growth projections and predicted that by 2018 a 1Gbit/s service downstream would be the norm (not a bad guess when I look at it with eight years of hindsight), and that we would need to find a new modulation technology to take advantage of the latent potential of a coaxial distribution medium.

Thankfully, the very smart folks in Europe working on ReDeSign project had already begun exploring the use of OFDM as a transmission medium and we were able to stand on the shoulders of giants to move things along rapidly. Looking back, it is amazing how quickly those involved in writing the DOCSIS 3.1 spec were able to get it completed, and I am equally amazed at how quickly the vendors have produced initial testable products ready for submission to CableLabs.

We are fortunate in working through CableLabs that we were able to bypass one of the most difficult parts of the innovation cycle, that being selling it to business partners as essential to maintain future services. Without their support, it is doubtful we would have product ready today.

The data conundrum
What I realized is we still do not fully understand the innovation cycle. Our businesses are based on data analysis for product development. How many can we sell? How much will it cost to create? How much to operate? To deploy? To maintain? These are questions that we expect our product teams to answer before they create any new products.

Does this approach work with innovation? I believe with incremental steps it is the logical approach. Whether adding more QAMs, viewer request channels, node splits, frequency changes, etc., it makes sense to run a business model that shows the ROI.

Want to learn more about where the cable industry is heading with next-gen technology? Then check out the agenda for our upcoming Cable Next-Gen Technologies & Strategies event, Tuesday, March 17, 2015, at The Cable Center in Denver.

But how do we apply that approach to step-function technologies? Moving from QAM-based DOCSIS to OFDM? It is much harder to study, due to the larger number of variables and volatility of technologies that are involved. The greater the variations, the less likelihood of accuracy.

Sometimes we need to move forward without analyzing excessive amounts of data. It is better to run small trials and experiments. Test a little, sell a little, learn a little. Repeat until such time that we either have proven or disproven the hypothesis.

Failure is not an option, but it is inevitable
As a left-brain person, I do not hesitate to take up new artistic endeavors. I never played the violin until I was 50 years old, but in the intervening years I have become proficient (or so I am told). The same goes for other artistic hobbies of mine, I don't worry about failing at something musical or arts-based because I realize that the only way to get better is to fail first.

The trick is to hedge your bets. Focus on the small gains and keep plugging away. It will get better with experience and by focusing on the end goal. You may not get there on the first try but failing early and often is fine, just try not to do it in full pubic view.

You will have to ask my wife about the early violin days…

Innovation disrupts from the inside out
Our businesses are based on a rhythmic pace. We get products into the shop for a lab trial, then a field trial, then back to the lab for integration into the whole, then back to the field, then finally deployment. Add product management, customer care, marketing, sales and field services to the mix. It all takes time.

I was fortunate to take an innovation boot camp led by CableLabs CEO Phil McKinney, where he talked about "corporate antibodies." Paraphrasing the section in his book, these antibodies attack anything that disrupts the rhythm and will try to shut it down. Early involvement with all boundary partners helps reduce the impact, but it still requires persistence to move these new technologies through. How many of us have heard the phrase "where are the business requirements" when looking at technologies that are three-plus years out?

It is at times best to take on the most far-reaching and complex problems and look for solutions. Not that you will reach all the way, but even if we only get halfway we are still much further than we would have been without trying at all. The distance from the start to the middle is much longer than from the middle to the end. It is easier to keep things moving once started than to get them started at all.

Not everything is a consensus
Very few of us enjoy conflict; we like things to be consensus-driven. However, there is a positive conflict that can be used to advance technology development. Finding those with similar though not completely alike ideas allows us to develop a better technology. Seek out those in opposition (the antibodies) with different ideas and learn from them. You will find that they have good ideas and if you can synthesize their thinking, your end technology will be much stronger.

Everything in its own time
As we learned with each successive DOCSIS technology wave, each step occurs in its own time. You make it up a mountain in small steps but not always in a straight line. We are taught early on that the shortest distance between two points is a straight line, but in real life there are no straight lines. We bob, weave, duck, sidestep, double back and eventually reach our destination.

Innovative technologies work the same way. We see where we need to get to (in the case of DOCSIS 3.1 it was reaching gigabit services), but we need to take on each challenge as we crest each hill. Putting it a different way, you need to be flexible and, when you move forward, there are times you have to step back or to the side and re-evaluate your approach. Adapt as needed and don't be afraid of looking like you are not 100% sure. Because there are times when we are not.

In a true team setting, you should feel comfortable with looking foolish in front of your teammates. Just ask my wife about the violin sometime…

— Jeff Finkelstein, Executive Director of Strategic Architecture, Cox Communications

About the Author(s)

Jeff Finkelstein

Jeff Finkelstein is the Executive Director of Network Architecture at Cox Communications in Atlanta, Georgia. He has been part of the engineering and architecture groups at Cox since 2002, and was part of the team responsible for the deployment of DOCSIS technologies, from DOCSIS 1.1 to DOCSIS 3.0. Jeff has made significant contributions to the access network design and deployments at Cox, and has the new role of being responsible for future technology planning of backbone, metro, edge, access, and home networks. He is now part of the CableLabs DOCSIS 3.1 PHY and MAC teams, in addition to being part of the IEEE EPOC specification effort.

Subscribe and receive the latest news from the industry.
Join 62,000+ members. Yes it's completely free.

You May Also Like