Hardware as a Service
New IT Business Models
Lately, I’ve encountered a lot about edge computing and the massification of IoT but, are our current cloud architecture patterns prepared for handling the velocity required for keeping the user experience as good as it is today?
Access to on-premises has been deprecated, leading to an increased dependency on IaaS for CD/CI; the construction of mini-data centers at multiple locations might become an entry barrier for small companies, and if we don’t deploy 5G at a good pace, even big corporations will be struck with bottlenecks and latency problems. That’s why it is important for future developments that we redesign our approach to Hardware procurement.
A recent article from McKinsey describes four paths for the IT infrastructure business models of OEM for the near future.
Software-led product or offer restructuring
Disaggregation of hardware appliances to their granular features and attributes like I/O, Latency, etcetera, to match them to the user’s needs. Added to an integrated control center to manage hardware-performance attributes and adjust underlying hardware or virtual resources as needed, and also an offer of several value-added software features like performance monitoring and administration, advanced analytics on usage, and proactive maintenance of underlying hardware.
Following recent trends and considering the possibility of future IoT solutions, the requirements for quantity and availability may overcome quality and longevity that are still the most alluring features for the average SME hardware consumer. A reduction in revenue margins might be compensated with a more diverse market segment and higher demand increasing sales, which might propel this shift to hyper-targeted products as we have seen in areas as ML.
Public-cloud and hybrid-cloud extensions
The position for an OEM should be to offer with their existing customers an extension of their on-premises value proposition in the cloud. Added to a creative buying program — call it subscriptions or whatever that is mainstream today ’cause the clients are more familiar with that model and are easy for both CFO to work with — that allows buyers to migrate their spending seamlessly between on-premises and cloud products. Driving a preference for the company’s cloud offerings over the competition as customers migrate their workloads to the cloud and their hardware. Also, they should enter into partnerships with one or more public-cloud providers to tap into a new customer segment of cloud-first buyers previously not exposed to those companies.
This is quite difficult to be done unless Microsoft or Google become an OEM — well there are some rumors about Google own chips architecture but whatever — , is not credible to believe that a cloud offering provides by HP or an alike company, might compete with AWS, unless is niche-focused; which is how new platforms are offered out there right now, but would that be scalable as a business model? Even if that’s the case, there’s no way of stopping Amazon or any other public cloud from mimicking possible targeted solutions and surpass the OEM cloud product offerings.
Only the idea about partnerships is a nice option in a short-term perspective, as many security and analytics firms are partnering with traditional cloud providers to leverage the actual infrastructure of customers with targeted cloud offerings, but even in these scenarios is difficult to see a business portfolio where both hardware appliances and owned cloud products have the same impact or at least both are relevant components from the main strategy of the enterprise.
There are some examples out there like Palo Alto and Azure that are worth taking note of, but unless both companies get to an agreement, the competitive advantage gained from these sort of transactions isn’t that big, and tying up some of your products to a specific cloud provider isn’t a great move, especially if you’re for example, from China.
I find it easier to believe if we turned to edge computing and more diverse hardware requirements, as public clouds won’t be able to keep up with that many necessities — even if they spend all their liquidity in buying startups — , and OEMs’ also won’t have the capacity to concurrently develop new hardware and software at the same time, these sort of teaming up might become more plausible, and lead to cooperation born from necessity instead of greed and anti-competitive practices, which could lead to more innovation.
Software-adjacent Portfolio Expansion
The companies that have made successful portfolio moves first, start by understanding their target decision-makers’ typical buying processes. Then they expand to other products that CTOs, CIOs, and other executives either already need or are likely to want in the future — basically, create your demand even before it exists — . Once launched, companies integrate these new products with their existing portfolio and create packages or bundles that holistically address those targeted customers’ needs.
This is an already proved strategy for any cloud-provider out there, they launch from time to time a new product or feature that you didn’t even know you needed but, once you get into their fantastic interface and see how many lines of code you might save from writing, and after reading how your productivity surge might also increase your revenue, you run to your boss to ask him to add it to the budget.
Dear McKinsey, this is just the fancy way of talking about horizontal integration and diversification, nothing new here…
Cloud-like subscription-pricing models
Companies that introduce software products — through one of the three approaches that we have reviewed — often offer the hardware on an up-front-payment model while increasing profitability through subsequent sales of software on a subscription model. Also, they can offer customers the option to pay on a per-use or consumption model. Companies that launch and scale hardware appliances successfully can make this pricing model’s economics viable by using contracting levers such as minimum contract commitments or cancellation fees.
Now, I see a contradiction with these. If our efforts are aligned with an agile strategy, and our clients and the market as well are demanding more personalization and flexibility, why I’m supposed to restrict my payments to timed-tie contracts, I get that in a subscription model is common to ask for payments every year — is also a bliss for the financial department — , but if all the innovation burst up and new startups are leading the competition, how is possible for a consumer to keep trying out different approaches to their physical infrastructure, are you telling me that in a trial period of two weeks he’s going to grasp how a firewall is going to protect all their hardware appliances for the rest of a year. Is quite discouraging to keep confronting these practices, unless the brand is already well-known, for a new business in the industry might prove pretty difficult to get those first contracts, that’s why you might want to take this suggestion with caution and try out some other payment models to secure a constant revenue.
It’s quite hard to propose a big change into an industry that’s already as consolidated as Hardware, and also due to the nature of the product, it doesn’t have that flexibility that software has shown during the last decade to produce new startups and solutions. If you review the four propositions, you might find that in reality, it does not have that much to do with the actual commodity and more with extra features or business functions. Is necessary to keep waiting to see if propositions like edge computing and the last market movements from giants like Nvidia, Intel, and AMD decide a new path for the IT Industry.
