When two clouds are better than one

We are living in a new era in retail, when IT and marketing need to do better than just “get along”. Traditional retailing methods have moved into a new world of social media, search, click & collect, email among others, and the real competition for your customer’s next order is no longer the shop next door, it’s the one in the cloud.

Working with retail clients, we can see the tension as IT has to accept that its “ownership of data” is no longer tenable, but then neither is marketing’s “ownership of the customer”. The reality is that both departments are battling with companies like Amazon, built from the ground up using cloud computing platforms. These new players have never differentiated between who owns data and customer insight – they are one and the same thing. And everyone in these retail operations collaborates within an architecture that supports that outlook

Retailers are wrestling with decisions about their data architecture, and cloud computing provides them access to levels of capacity, functionality and inter-operability that were unheard of a few years ago:


Your access to the cloud is pretty much limitless. This makes it cheap, but it means much more than cost-cutting. Endless amounts of space changes our whole approach to problem-solving. Instead of creating a repository to which everything needs to connect, we now build (potentially) limitless numbers of tables that can be updated in real time – the only restriction being the efficiency with which your other tools can feed it.


So what? Most retailers will tell you they already have too much data.

The abilities of machine learning to tidy up messy datasets, run thousands of calculations and serve up insights are in their infancy, but this is the game changer. There’s no such thing as too much data – the problem is not having enough time to look and act on it. Marketers in retail need to learn IT-type skills in building numerous tests that increasingly guide them to understand which offers work best, whilst never losing their flair and creativity. Whilst IT needs to learn how best to support the business of “delighting the customer” with an IT operation that is driven by customer signals, not just product and stock information.


If capacity is an enabler, and functionality is the game changer, what’s so great about inter-operability?

The answer is activation. Where we used to come up with a campaign idea and then go find the customers to market it to, we now have the direct links between the insight and the next best action built into the same ecosystem. For example, if a valued customer has not opened an email we’ve sent them, what should we do? If a prospect finds our site through a highly valued search term and fills a basket, but doesn’t buy, can we send her an automated signal? When an existing customer searches for our opening hours on a site, is there a way we can help him in that moment of need?

Traditionally, all the scenarios above require painstaking planning and coordination between departments, but cloud-based orchestration makes this a far more seamless and efficient process. And we can test a range of test scenarios to see which one is the most efficient and effective and improve performance over time.

So, why would I need two clouds?

A common misconception is the idea that a company needs to choose only one cloud partner. This mindset is largely based on our historical practice of running beauty parades that compare features, followed by excoriating price negotiations with suppliers. But these types of feature and cost comparisons can be misleading. For one thing, the hardware costs are no longer really a factor: after the set-up phase – so long as the project is well planned – the expense will mainly be the human cost of hiring and training experts to manage the data.

The key step is for IT and marketing to genuinely align on each other’s requirements and build the right architecture to support their use cases. This is often only possible by using more than one cloud solution. One example of the need for multiple cloud partners is in the access to data. Google provides efficient access to its marketing data on GCP, it can be recreated on another cloud, but the underlying cost of this operation is by far higher than the integration mechanisms which enable one cloud platform to become an efficient data provider for the other. This is one example, and there are others.

The critical step is for a full assessment of the required use cases, and expert support in specifying the deployment of the data. Ultimately, IT and marketing need to collaborate to fully enable the next generation of customer engagement. And they need to build their tech stack together, to ensure that they focus on delivering the best service to the customer in the most efficient way.