I’ve been writing for the last couple of weeks about Omni-Channel architecture (part 1 and part 2), as presented by Anoop Kulshreshtha of Vitamin Shoppe, along with my own thoughts on the topic. This week is part 3, the last installment in the series.

To recap briefly, the elements I covered earlier were a single view of the customer and order management in part 1, and product information management (PIM), and enterprise content and digital asset management (DAM) in part 2. This week I’ll cover the nitty-gritty IT stuff (yes, I don’t include those other topics as nitty-gritty IT, though you could argue the point): service-oriented architecture (SOA) and an omnichannel-aligned IT strategy. Plus the bonus topic that I think Kulshreshtha missed.

Service Oriented Architecture (SOA)

Kulshreshtha made the case that the heart of any omnichannel strategy needs to reside with a service oriented architecture (SOA), and message control is the heart of SOA. No point-to-point interfaces — an IT person’s ideal end-state for the enterprise.

It’s interesting to hear this argument made in the context of omnichannel. Usually you hear it as just plain good sense for the business. However, business users tend to view SOA as IT for IT’s sake and they historically have not wanted to pay for that. Services make good sense in omnichannel — even more so than the other areas that have historically been stood up as prime examples of where SOA works best — typically when having to interact or integrate with external parties, like in the supply chain.

Omnichannel has an even stronger case to make for using SOA. Channel-agnostic integrations become critical, especially when whatever innovative offerings you put on the eCommerce platform need to make it to the Facebook page and the mobile site — and get pushed out to third party affiliates or search engines. I have to wonder if Google’s product search has done more for data standardization than all of the digital catalog efforts to come before it combined.

The challenge, as always, with SOA, is how to get the business to pay for it. Kulshreshtha recommends a piece-meal approach, where 6-month sized chunks deliver short-term wins in the context of a long-term strategy. However, in order to do that successfully, you need to have a long-term strategy, which is where Kulshreshtha’s next omni-channel architecture component comes into play.

Omnichannel-aligned IT Strategy

This is where all the other elements come together. Kulshreshtha argues that this is also the hardest part — it requires discipline and a shared vision. He recommends that no IT solution or implementation be project-specific — that it should be validated against a 5-year omnichannel plan. And anything that deals with data needs particular attention — any data that is enterprise class should be treated with respect. In order to get there, though, sometimes IT has to be a little stealthy. When asked about the challenge of selling architecture to the business, he responded that IT people should never sell IT projects. They should always sell business value. For example, don’t try to sell the need for an order management system. Instead sell the idea of inventory visibility in 6 months, click & collect in another 6 months, endless aisle in another 6 months, and finally full-on OMS at the end of 24 months.

So What’s Missing?

As I sat through Kulshreshtha’s fantastic presentation, it occurred to me that he might be missing one item from his list. While everything on here I agree is needed, he never addressed the front-end of the customer experience. If the goal is an omnichannel experience for customers, how does a retailer ensure that the experience they present is both consistent across channels and designed to make the most of each channel? Where is a single customer interaction platform on his list?

It may be he left it off because such a thing doesn’t really exist. I certainly got a lot of perspective on that from some retailers at the event. There are a lot of people advertising the idea of a customer interaction platform that crosses channels, but no one even close to delivering on that yet. In some cases, it’s slideware promises, in other cases, the devil is in the details. And for retailers trying to resolve the future of eCommerce, POS, mobile, and social, it’s a lot of frustration. They have right-now problems, and not a lot of solutions.

From our benchmark research, we find retailers mixed or outright contradictory about the idea — they say that “a single customer interaction platform” is the number one thing on their buying list right now, but they also say that they envision a future where a single platform maybe provides a lot of the backend stuff — the PIM and CRM and DAM that Kulshreshtha listed off — and leaves the front-end interaction stuff specific to each channel.

For my part, it comes back to SOA. How much of the interaction capabilities need to be abstracted away from the channel vs. left specific to the channel? Are we talking 50% common across channels? 80%? If so much can be pulled out – pricing, product content, validating a credit card, confirming a transaction, a shopping cart, a wish list, loyalty points, etc. — then what’s left? A presentation layer? A few things that are channel specific, like maybe promotions (which I would argue could still probably be managed as a common service across channels even if your promotions are channel-specific)?

It’s disappointing to me to see vendors who have all of the pieces of the puzzle — eCommerce, POS, contact center — or even those that have pieces and partnerships for the rest — and still don’t have a comprehensive offering for retailers on the customer interaction front.

It looks like that piece of the omnichannel architecture is still waiting to be invented.