Is digital performance getting too cloudy?
According to a global study from Dynatrace, 76 percent of the of 800 CIOs surveyed think the complexity caused by modern, cloud-centric ecosystems could soon make it impossible to manage digital performance efficiently.
The survey found that a single web or mobile transaction now crosses an average of 35 different technology systems or components, compared to 22 just five years ago.
Fifty-three percent of CIOs plan to deploy more technologies in the next 12 months, and the technologies they plan to adopt — multi-cloud (95 percent), microservices (88 percent) and containers (86 percent) — are likely to heighten complexities:
- Seventy-six percent of CIOs say multi-cloud makes it especially difficult and time-consuming to monitor and understand the impact that cloud services have on the user-experience;
- Seventy-two percent are frustrated that IT has to spend so much time setting-up monitoring for different cloud environments when deploying new services;
- Seventy-two percent say monitoring the performance of microservices in real-time is almost impossible;
- Maintaining and configuring performance monitoring (56 percent) and identifying service dependencies and interactions (54 percent) are the top challenges CIOs identify with managing microservices and containers.
As a result of the mounting complexity, IT teams now spend an average of 29 percent of their time dealing with digital performance problems.
As CIOs search for a solution to these challenges, four in five (81 percent) CIOs said they think artificial intelligence (AI) will be critical to IT’s ability to master increasing complexity. Eighty-three percent are either already deploying or planning to deploy AI in the next 12 months.
“You can’t rely on humans to synthesize and analyze data anymore, nor a bag of independent tools,” said Matthias Scharer, VP of business operations, Dynatrace, in a statement. “You need to be able to auto detect and instrument these environments in real time, and most importantly use AI to pinpoint problems with precision and set your environment on a path of auto-remediation to ensure optimal performance and experience from an end users’ perspective.”
Discussion Questions
DISCUSSION QUESTIONS: Do you see major hurdles for retail IT teams in transitioning to multi-cloud ecosystems and cloud native architecture? Is artificial intelligence the answer to managing the hyper-complexity or does it carry its own risks?
Poll
BrainTrust
Paula Rosenblum
Co-founder, RSR Research
Ralph Jacobson
Global Retail & CPG Sales Strategist, IBM
Isn’t this a “rich person dilemma?” Almost a decade ago, it was clear that SaaS (cloud-based) technology was the far less expensive means of running sizeable retail and IT technologies. The world gravitated this way. Yes the number of systems has increased, but so have the capabilities and specialties of each. Even with the challenge of multiple systems abound, IT teams are providing far more value to their organization, and customers by running multi-tenant, cloud-specialized technologies. The part to consider is which systems to select as core technologies that the IT team purchases as a best-of-breed point solution to integrate into their architecture. Select poorly, and as an example you end up with a unified commerce solution that does some things well — ERP — and bombs on other aspects that the vendor isn’t as good at, like the OMS.
The key to all things cloud is composability — IT teams that understand this and have some understanding of the “do’s” and “dont’s” of multi-cloud system architecture will have success. This is going to require a heavy investment in resources (people, capital, time) and buy-in from the C-suite.
AI will be essential to understanding the technology and identifying bugs within the complex ecosystem otherwise the IT team might never see the sun again.
Just as companies have sprouted to monitor and actually balance the use of resources in cloud computing (e.g. Turbonomic), companies will be formed to provide products that manage these complex and highly-variable tasks on behalf of IT departments. It is clearly not efficient (or maybe even possible) for people to do this on their own. As has always been the case, a move in the direction of a new technology begets yet another opportunity for others to provide a solution. And the consumer will never know it’s happening.
As we move increasingly to multi-cloud ecosystems and leveraging AI in general, there are major hurdles and learning curves for retail IT organizations to overcome. While harnessing the power of AI has significant benefits, there are some resourcing risks, as IT organizations will need to source the right talent and perhaps create new roles in their organizations to embrace these changes.
Retail organizations as it stands are already challenged to run the day-to-day operations, and invest their time into upgrades and other technology transformation. Once you introduce the complexities of new multi-cloud architectures and AI technologies, you may create a situation that is too overwhelming for your organization to take on.
In these scenarios, retail IT departments may choose to outsource or collaborate with AI/cloud consulting organizations to take a crawl/walk/run approach with these new technologies. There is also an element of change management and culture change that comes along with these new technologies.
Right on point Brandon. My only quibble is that your last sentence should not be almost a throw-away line — it is THE most mission-critical issue. We see this in the pathetic record of M&A and now we’re seeing it in the obsession with deep data that we do not know how to use to any practical outcome. No matter how clever we think we are. It all boils down to how people relate to each other.
I couldn’t agree more Ian!
The people component is perhaps the most critical area that gets overlooked by organizations. Without the trust, transparency and buy-in no technology transformation will be successful. There’s the classic people, process and technology winning combination with any innovation. You simply can’t underscore the importance of how people work, trust and collaborate with each other.
This is an interesting question around unintended consequences. Microservices do offer a lot of promise — the true “on-demand” solution at a very granular level. But there’s no doubt that tracing tools have to become more intelligent to help sort through where bottlenecks are.
Bottlenecks have always been an issue with the web, with or without cloud services — but the problem is exacerbated by microservices. But they’re worth it, I think. We just need better tools.
Part of the problem with technology is the challenge of what it “should” do and what it “can” do. The truth is that we are still years away from much of the technology introduced today being perfect, and there is the risk of what we’re are investing in today will be obsolete before it is fully developed. So IT departments, CIO’s and ALL retailers have to be careful with what they spend their money on and how it will impact their business, and take it one step at a time.
Unfortunately, too many retailers are dropping the ball on the essential daily needs of their business — like store personnel, store associate training, store remodels and better merchandising. Is the customer going to spend more money when the well-trained associate makes them feel special, makes the right recommendations and satisfies them? Or will the customer spend more money because they heard about all the technology the company is planning to introduce? The answer is an obvious one, so yes we need technology and it’s the future, but be sure you invest in it carefully and make sure it will be right for your business.
Multi-cloud ecosystems are a consequence of specialization for specific computing and storage applications. Developers have created optimized solutions for delivering a lower-cost or specialized function. While this approach provides better solutions at a lower cost, the implication is that retailer IT departments need to integrate these multiple clouds into a solution that works best for the retailer.
Companies have stepped up to help retailers manage and control this integration using more AI to help manage the ecosystem created. So retailers who have the luxury to pick and choose the clouds that work best for them, with the resulting complexity, also have a choice of multi-cloud management offerings to manage and control the result. These AI tools will get better over time.
The quote in the final paragraph is gobblygook! With statements like that, no wonder IT is a quagmire of complexity!
But seriously, one of the issues is that retailers — largely because of operational and financial necessity — tend to upgrade and invest in different systems at different times. This piecemeal approach does create complexity and challenges, especially when going in a new direction by using cloud native architecture.
The challenge is not just technical; it’s also cultural. Most retail IT teams are used to working traditionally — a single block of executable code which they can monitor and update. Newer ways of operating change that culture. For example, more people are needed to check code earlier in the process, and working is more fragmented as different blocks of code can be updated independently.
These are traditional retailers will make, mainly because these new ecosystems are much more flexible and advantageous. However, most retailers are still in the transition phase, which is painful.
You live by the cloud — you die by the cloud.
We are so clever at analyzing data that data is now analyzing us. We can’t handle it so we turn to the oxymoron of Artificial Intelligence, resigning our humanness to devices which we hope will rescue us and make sense of our world.
In the retail world all of this craziness and expense is to accomplish what exactly? To figure out how to better serve our customers? Is it just me … or is there a tragic disconnect and abdication here?
I have a different take on this question. AI in this context is about analytics to monitor and improve the efficiency of multi-cloud environments. I think the challenge has as much to do with balancing the advantages of best-of-breed clouds with cloud customer experience platforms that account for many business processes (and have therefore integration efficiencies built in). AI then becomes secret sauce to inform those business processes in differentiated ways — which retailers absolutely require if they hope to realize the reported benefits of digital transformation. I called it the coming wave of efficient effectiveness back in October of last year.
I think that there is a real opportunity for consultants to come to retailers and get down to the basics of determining precisely what they need to know and get that information for them. A 30,000 foot view can focus on selecting systems that deliver and minimize complexity. However, as technologies and software rapidly change one challenge is to maintain that 30,000 foot view.
I think it’s likely that organizations are adding on and adding on to their systems instead of revisiting the basics from time to time, almost viewing needs and solutions as though there was nothing on that white board.
After looking through some great comments on this article, I see this phenomenon as potentially a “ready, fire, aim” situation. Retailers heard that “everyone’s moving to the cloud,” so most were compelled to act quickly. Now, AI is the flavor of the month. On the horizon is quantum computing. The trouble is, rarely are implementation strategies developed with specific metrics defined along with the methods to capture those metrics.
I don’t so much see the multi-cloud environments causing the performance measurement challenges. I see the relevant data not being captured nor analyzed in the first place. Now, with AI, certainly there are systems that can uncover the “dark data” that legacy systems cannot see. But, once again, retailers need to define exactly what it is they want to measure, so the AI systems can leverage their machine learning capabilities with human guidance.
That’s the point, as was highlighted by Ian and Brandon in their commentary, that human planning and execution is the most important component to effective cloud, AI or any other technology deployment.
Data integration (and quality) is not a new problem for retail and CPG companies … both on and off-premise.
The major hurdle is, in two years this will be outdated.
Retailers have run into a couple of issues around cloud, unintentionally. One is the nature of how IT systems have evolved and been upgraded by retailers over the years — a very piecemeal process that results in significant complexity over time. Second, the real benefit of the cloud has materialized not in cost savings or ease-of-use, but in productivity gains and scalability. While retail CIOs may have originally looked to the cloud for savings, what they really achieved was increased scale and capability they couldn’t otherwise achieve in-house. The result? A different kind of complexity — one that required new skill sets their staff was not used to. This isn’t the same type of complexity they were trying to eliminate, but it certainly didn’t simplify things as expected.
Then you have the question of how “all in” did they go with the cloud? I know a few retailers who went “all in” and moved all of their applications to the cloud and did achieve those savings and simplified operations they wanted, but it was a massive effort over time to get there! These were specialty retailers of modest size and I can imagine larger retailers having a greater challenge to migrate.
One area not mentioned in the article that also contributes to the complexity is the retailer’s network infrastructure. Many times I’ve seen this ignored until it’s too late and performance suffers. Of course, that means the benefits are not realized in the desired (and expected) timeline which makes the cloud migration seem even more complex! The right answer is to plan for this early in the process and ensure you not only address capacity issues but also look at implementing tools that deliver visibility into the application traffic you’re sending and receiving from the cloud along with the ability to control the performance of those apps. Without those tools, the cloud will continue to be an unrealized promise!
The benefits of multi-cloud and native cloud architectures are enormous. While it does add some complexity from a system monitoring and performance optimization perspective, it is worth it.
Leveraging AI and other system analysis tools will continue to make this process easier. This will be a great opportunity for an entrepreneurial company to automate these processes with innovative solutions and services.