Bimodal: IT innovation or excuse?
Can you create with bimodal IT an adaptive, fluid enterprise that both anticipates changing demands and evolves to meet these?
The IT industry has heard a lot of talk recently about the concept of “Bimodal IT” originated by Gartner Research. Gartner Research states that “forty-five percent of CIOs state they currently have a fast mode of operation” and predicts that “75% of IT organizations will be bi-modal in some way by 2017”, so soon we will be well along the adoption path.
You can read the official Gartner definition in their glossary, but I believe it posits that IT should be split into two separate groups. My take is that the first, “Mode 1”, is the traditional IT with which many are familiar, focused on “run safe”, where predictability and performance are key. Think of core banking transactional systems, sensitive customer databases and enterprise heavyweight ERP. “Mode 2”, however, concentrates on business agility, and the need to frequently deliver innovation, leveraging concepts such as Agile and DevOps. Think cool mobile apps and omnichannel initiatives designed to offer a seamless customer experience.
I attended the recent Gartner AADI summit in London and viewed Gartner as being still very much behind the Bimodal IT approach, with several presentations/discussions around it. The concept in many ways makes sense. After all, you can’t force Agile rates of change on a heavyweight IT infrastructure that wasn’t designed to be updated weekly, daily or even in some cases hourly? Nor does it have to. There are structural, cultural and technical reasons for not taking the same view across the entire IT organization.
Another Gartner viewpoint comes to mind—that of a “Pace Layer” strategy; organizations have systems of record, differentiation and innovation, with the first changing less often and using traditional approaches such as Waterfall, and the latter the most dynamic and Agile-focused. The London conference revealed that the reality can at times be more fluid, and at others, patchier. You can find examples of Agile adoption in certain systems of record, and equally signs of Mode 1 thinking in innovation systems. I like this acceptance of the fact that (like many things in life) the reality can be less uniform, less black and white and not so clearly defined. A recent Oxford Economics paper sponsored by CA, for example, found evidence that more established departments such as HR were actually ahead of cutting-edge ones such Marketing in adopting new development processes and tools.
But Bimodal also has its detractors, one of the most vocal being analyst Jason Bloomberg of Intellyx. Last October Bloomberg penned a compelling blog that outlined the case against Bimodal IT, with Bloomberg going so far as to say, “Bimodal is simply an excuse to keep doing IT poorly”, and stating that it perpetuates siloed behavior. Bloomberg makes the very valid point that Agile is about small, customer-centric changes rather than a big bang launch that risks major systems collapsing. In essence, established IT doesn’t need to be afraid of innovation if it is staged, incremental and can address defects rapidly with a quick update.
One of the most interesting takes on the concept has been from Simon Wardley, who is also skeptical of Bimodal IT, as it groups IT into either “pioneers” doing the frequently changing, chaotic yet exciting cool stuff and “town planners” who take a “steady as we go” approach and do the dull, orderly yet essential things. He points out that this model misses a crucial group, the “settlers” who take the innovative approaches and tools initiated by the pioneers and develops them to a mature stage at which point the town planners are willing to industrialize these, ensure scalability and make them part of the furniture.
If I was explaining the two Bimodal descriptions at an IT conference and I asked “Who would prefer to be a member of a Mode 2 team?”, I estimate more than 90 percent of hands would go up, with 10 percent or less opting for Mode 1 when asked. This is goes at the heart of the Bimodal IT issue, as recruitment of the smartest and most dynamic brains into Mode 1 could well be a challenge going forward. The developer community is abuzz with debates around open source, DevOps, Agile, containers, APIs and more—basically, there’s a lot happening over in Mode 1 world. How do operations folks entice genuine talent who want to make their company’s Mode 1 the best Mode 1 around? The established Mode 1 teams may not need new hires right now, but ultimately they will retire and who takes their place?
There’s also the relationship between the two groups to address. Do Mode 2 folks look down on those in Mode 1 teams? Do Mode 1 employees feel their role is unappreciated now the new kids on the block are receiving all the kudos? DevOps can certainly help here, with its focus on collaboration, shared goals and greater transparency.
One topic I was strongly in agreement with Gartner on at the AADI event was the significance of APIs, the “building blocks of the application economy” as we like to call them at CA Technologies. The way in which APIs are exposed/shared, updated and managed overall will have a huge impact on an organization’s ability to enter new markets, acquire new business partners and bring new products and services to market. Take Ice Mobile for example, who provide a series of personalized offers and promotions delivered via a shopper’s smartphone. In combination with technology partner Enable-U, Ice Mobile use CA API Management to fast track their integrations with retailers and in doing so increase the chances of winning a new customer as they can instantly address the “Like the idea, but how will it work with our complex IT infrastructure?” initial pushback. This has also cut down the typical time-to-market from 14 weeks down to eight weeks. It’s much easier to expand adoption of a great digital loyalty program when the customer (retailer) sees the benefits faster.
If we could be like Marty McFly and flash forward in a time machine DeLorean to 2018’s enterprise IT organization, I would be keen to ask them (along with who won the last three Superbowls), what percentage of your team would you label Mode 1, and what percentage Mode 2? How has recruitment into Mode 1 teams fared versus Mode 2? What are the salary levels and annual incremental raises afforded those in each Mode? How motivated, energized and engaged are the Mode 1 teams compared to Mode 2? What cross-pollination exists between these two resource pools?
One of the debates around Bimodal IT is around it being a desired end state or simply a transitional one on the path to an ultimately Mode 2-centric architecture. In either case, I see the role of “settlers” as critical, and taking the pivotal role of taking an example of innovation in one project, program or line-of-business and ensuring that the wider entity benefits. DevOps can have a significant role to play here, both as connective tissue between Mode 1 and Mode 2 groups, but also in demonstrating that the Mode 1 innovation does not need to be at the risk of degraded core systems, outages and governance issues. This in turn fits into part of a wider digital transformation strategy that creates an adaptive, fluid enterprise that both anticipates changing demands and evolves to meet these.