Enterprise architecture at Infosys works at the intersection of business and technology to deliver tangible business outcomes and value in a timely manner by leveraging architecture and technology innovatively, extensively, and at optimal costs. Stay up-to-date on the latest trends and discussions in architecture, business capabilities, digital, cloud, application programming interfaces (APIs), innovations, and more.

« January 2019 | Main | September 2019 »

May 7, 2019

How to Evolve an Aging eCommerce Platform: Suggestions to Frame a Way Forward

By Pierluca Riminucci, Sr. Principal Enterprise Strategic Architect

Many companies in the retail sector are faced with a common dilemma about how to evolve their existing and now aging eCommerce platforms.  This is a technical but also a business key decision that is more often faced by established "brick & mortar" retail companies.  Such traditional retail businesses have typically embraced eCommerce a decade or so ago, slowly building a capability, predominantly in a tactical way, perhaps more as followers.

In my opinion, a good way to tackle this far reaching strategic decision, is to move from one fundamental truth: the future has in store a lot of difficult-to-predict disruptions and changes.  If we truly believe that statement, only seemingly hollow, then we are in a better position to start uncovering its less obvious consequences, the main of which implies that the real end goal should be building an innovation-enabling technical foundation, rather than simply selecting another eCommerce platform.

And the reason is simple: one of the truly key strategic business advantage of companies will be their ability to evolve their business models quickly, reliably and effectively, and with minimum additional investments.

That will probably be the game of the next few years ahead of us which will determine the winners and the losers.

Now hopefully is more clear why, to make an informed decision regarding "what to do with my current eCommerce platform", is beneficial to broaden the perspective and start addressing - or at least framing in the background - the underlying deeper questions revolving at establishing what is my company vision for the future in the space of "digital retail".  However, it is often very difficult for a Company to articulate a sharp, coherent and concrete-enough-to-be executable vision able to shed light into the depths of uncertainty brought in by a rapidly approaching technology-led disruption.

And that brings us back to the earlier foundational opening statement: The future has in store a lot of uncertainty.  However, amidst uncertainty, there is also a recognizable enough common trend, i.e. the former sharp distinction between eCommerce and "Commerce" is getting less and less meaningful due to technology advances.

Experience in store will likely be augmented by an ever increasing digital dimension.  Products will be "experienced" by Virtual Reality (VR) headset[1] or garments tried-on both physically but also digitally[2].  Store visits will be enriched by virtual aisles technologies, garments could be tried-on in store to see how they fit, 3D pictures could be made and then "saved for later" allowing that garment to be bought on-line with confidence, and at the same time helping relieve the notorious eCommerce high return problem. Digital information points (or "hooks") about products or life-style associated with products could be disseminated in the store offering customers a truly full immersion experience.   

At the same time, the on-line channel can be enhanced by the possibility of leveraging the capabilities that an "brick-and-mortar" infrastructure offers.  For instance, body measurements could be scanned in store by special purpose equipment to enable an unprecedented accuracy in matching garment wearability to potential customer's unique body shapes. This in turn will open up a raft of new and more powerful ways to provide product recommendations, like sending customers "virtual pictures" of them wearing a new garment and even open-up new quasi-mass-market-made-to-measure manufacturing approaches.

I am aware that many of these technologies are not quite ready yet, however there is enough happening to be persuaded that the winners will be those who will be able to quickly and seamlessly integrate new pieces of innovative and best-in-class technologies in their technical ecosystem to enable rapid business models innovation.

In the light of all that, the initial question "what to do with my aging eCommerce platform" takes on a distinctive different perspective. It is not a matter anymore of simply comparing pro-and-cons of different vendors' eCommerce platforms but more appropriately identifying the key quality attributes of an enterprise architecture that would allow a seamless integration (between eCommerce and Commerce operations) along with the possibility to be rapidly evolved into new unknown directions.

So hopefully by now it should be clearer that a more effective way to frame the initial question would be by first identifying a set of guiding architecture principles and then properly spelling out their key consequences to build a robust and far reaching decision framework.

In the remaining part of this article I will illustrate what these architecture principles and their consequences might look like, as a title of example. Obviously, I am aware that the specifics of each company situation might require some fine tuning, given that a fit-for-all solution is unlikely to be really effective.

A guiding set of architecture principles for evolving your eCommerce platform.

AP 1.  Architecture decisions should aim at minimizing technical constraints:

·         Consequence 1: Carefully trade-off the convenience of out-of-the-box comprehensive - but also constraining - eCommerce platform with future yet-unknown requirements.

·         Consequence 2: Watch-out for vendor lock-in warning flags (e.g. proprietary interfaces, lack of extensibility, hegemonic vendor strategy, proprietary languages, etc.).

·         Consequence 3: Take control of your integration strategy and enabling interfaces (use open standard to hide product proprietary protocols).

·         Consequence 4: Carefully look at the extensibility and integration mechanisms of each eCommerce platform (e.g. prefer mainstream mechanisms).

AP 2. Reduce execution friction to innovation:

·         Consequence 1: Aim at achieving a loosely coupled architecture.

·         Consequence 2: Preserve the ability to easily plug & play specialised modules (COTS or bespoke) that you might need to acquire or develop in future.

·         Consequence 3: Institute a lean and proactive architecture governance function to avoid architecture degradation.

·         Consequence 4: Preserve a strong semantic in terms of each module' remit but also in terms of overall coherency and clearly defined interfaces.

·         Consequence 5: Carefully evaluate the needs of introducing new technology paradigm against the increase in fragmentation of required skills (stay mainstream).

AP3. Get more for less and reduce costs (TOC & OPEX):

·         Consequence 1: Leverage existing investments whereas possible. There are no silver bullet platforms that alone can solve your business strategy execution problems.

·         Consequence 2: Constantly maintain your enterprise architecture aligned to a semantically strong domain model (refactor it each time it is feasible during transformation programmes).

·         Consequence 3: Consider Open Source platforms since they tend to be less "hegemonic" and more interoperable.

·         Consequence 4: Consider adopting a cloud strategy but only after a careful analysis of benefits and related economics.

AP4: Think in terms of Omni-channel architecture rather than eCommerce platform:

·         Consequence 1: eCommerce function is not likely to stand alone anymore in the future.

·         Consequence 2: Unify all your commerce function enabling capabilities whilst providing well-defined variation points to cater for channel flexibility.

·         Consequence 3: Strive to build an end-to-end coherent data flow as the main fabric to achieve end to end seamless business integration.

·         Consequence 4: Preserve your ability to leverage eCommerce capabilities to digitized brick & mortar experience.

·         Consequence 5: Collect customer data through each channel and pull it all together to execute a mutually reinforcing cross-channel communication strategy.


In this article I have articulated the need of adopting a holistic view about eCommerce and Commerce capabilities when considering how to evolve your existing legacy eCommerce platform.  I have presented a set of guiding architecture principles that could help framing the decision. The key point behind such principles revolves around the need of preserving your ability to stay in control since your company is likely to need to evolve and rapidly introduce innovation to simply stay still in the market place.


Inder Sidhu with T.C. Doyle, The Digital Revolution. Pearson 2016

[1] Indeed, this is already happening for instance with Audi. Cf. Inder Sidhu et al. "The Digital Revolution" (2016)

[2] This is what "magic mirror" technology is promising to deliver, though not quite there yet in terms of realism.

Reducing Disruptions in Manufacturing using the Internet of Things

By Sivan Veera, Sr. Principal Enterprise Strategic Architect

Shop floors have evolved from simple job-card-based push scheduling to advanced Kanban based pull scheduling. However, even the most sophisticated Kanban systems suffer from various interruptions. A technique called fluid scheduling leverages the Internet of Things (IoT) to solve this problem.  Fluid scheduling is a fundamental change where scheduling happens by what parts are currently available rather than what should be available based on a predetermined schedule. IoT based fluid shop floors also move scheduling out of the shop floor to centralized command centers where multiple physical shop floors can work as a single logical shop floor. This aggregation can plug-in multiple shop floors together in a "shop floor as a service" model creating a pathway for disruptive business models.

What is a Fluid Shop floor and how does IoT help?

Figure 1 illustrates an IoT based Fluid Shop floor. Sensors are mounted on pallets and count parts inside the pallet in regular intervals. Sensors stream these counts to a Central Scheduling Center which are external to the shop floor. Scheduling centers are networked to machines and they send machine schedules in near real time. Scheduling centers consider the availability of parts in real time, continuously update schedules and push the schedules to the machines. Scheduling centers also move parts from pallet to pallet based on the machine schedule. This type of scheduling is different from Kanban based scheduling where parts required are computed for each machine and updated with the assumption that these parts will arrive at right time at right machine at right quantity. This type of IoT based situational scheduling is also different from conventional push scheduling where parts are scheduled for a long period say a week or month and are not updated in between based on disruptions.


Figure 1: IoT based Fluid Shop floor

Role of IoT in Fluid Shop Floors

IoT sensors mounted on every pallet count the parts inside them and stream the data to Central Scheduling Centers. By knowing the exact parts available at each machine and also the parts available in the incoming area of the shop floor itself, scheduling centers will be able to calculate schedules in real time and update the machines. Since schedules need to be calculated real time and they are compute resource intensive, cloud computing fits well for scheduling centers. When scheduling centers integrate across many shop floors, high velocity ingestion from multiple sensors can also be handled effectively by cloud computing. For the sensors to identify parts in the pallet, each part in the shop floor need to be tagged with RFID tags. Sensors stream the part identification and pallet information to the scheduling centers.

What Business Models can IoT Fluid Shop Floors disrupt?

In today's business model, large shop floors are built around producing a small set of Products. Shop floors are optimized for mass production. Introducing new products in market takes years of building custom designed shop floors. Fluid scheduling lifts this limitation.  Many IoT based fluid shop floors can be assembled together into one logical shop floor and produce different parts in a short time. Many startups can build platforms to assemble these shop floors together in a short time. New companies will emerge building and launching products in a shorter time disrupting large established corporations. Pricing of products also will significantly change since large investments need not be made on large shop floors.

What can we expect in the future?

The Internet of Things enables shop floors to transform to the next level of agility and sophistication by dynamically modifying machine schedules based on the available of parts at that moment. This eliminates waste even beyond the gains made by Kanban based scheduling. With IoT based fluid shop floors, customers will be able to order parts by supplying designs and having those parts delivered within weeks with less upfront cost. Will there be a market place where multiple shop floors come and go as a "shop floor as a service?" Can customers bring new products to the market every quarter with such a service? Can shop floors be completely automated and controlled externally?

With the advent of IoT enabled fluid shop floors, we may witness disruption in business models at a rate we have not seen since the advent of industrial revolution.  What do you think?


Subscribe to this blog's feed

Follow us on

Blogger Profiles

Infosys on Twitter