There’s an old story that illustrates the problems businesses create for themselves when their various units operate as silos. Appropriately, it’s about a farmer.
As the story goes, there once was a farmer in Iowa who, in a single week, got more than 200 emails from Microsoft. Some came from the Windows group, others from the Office group. Some offered a deal to sign up as a new customer. Others thanked him for his continued loyalty as a “valued” customer.
One Customer, Multiple Business Units
Each group at Microsoft saw the farmer as its own individual customer. In their discrete efforts to reach out to their own business targets, the various business units collectively inundated the farmer with messages. In addition to causing a good deal of confusion, their duplicative efforts wasted money and had little chance of yielding positive results.
Thankfully, this story first started making the rounds more than a decade ago. Microsoft has since made changes to prevent something like that from happening again.
So Many Segments, So Little Time
In that story, one farmer represented multiple segments to Microsoft. For some groups, he was a customer, for others he was a prospect. Besides the obvious silos between product groups like Windows, Office and Servers, there were also the silos that had evolved around the company’s customer segments. The farmer “belonged” to many different groups. His name was included on lists in multiple categories, such as agriculture, small business, central district, direct consumer, third-party OEM partner customer, and industry association member. And as an actual customer, he “belonged” to the finance group that managed his account.
From the farmer’s perspective, Microsoft looked like a single company. He expected every group across the company to recognize him and to treat him as an individual. He did not think of himself as a Windows customer, an Office customer, a business account or a consumer. He was a Microsoft customer.
I wish this were an isolated example of different marketing groups within a company living in their own silos, with each claiming “ownership” of the same customers. Unfortunately, having worked with clients from Fortune 500 companies for decades, I have to say that scenarios like that are the norm in large enterprises.
Moving Forward With Data Management
Developing a unified customer experience relies on these two primary components:
- Organizational readiness: Developing a plan to solve the problem of separate groups “owning” the same customers, with each of them handling those customers independently from within own silos.
- Data management: Building and maintaining multiple, overlapping customer records and profiles, each independent of the data contained in other silos.
For the purposes of this column, I will focus on data. In a subsequent column, I will tackle the structural issues that are tied deeply to the way businesses are organized, how people are measured and how customer-centric a company really is, all of which bubble up to organizational readiness.
Over time, companies build operational databases, deploy customer relationship management (CRM) systems and set up multiple marketing technology systems, including marketing resource management (MRM) systems, data management platforms (DMP), campaign management tools, email systems, social marketing tools and customer service systems. Multiply the number of systems by the number of markets, countries and/or divisions within a large enterprise and you can easily see how data can become siloed.
Recently, I was working with a company that “discovered” it was using 11 different email systems, each with its own customer database and communication programs, and each requiring support, management and expertise. This scenario is prevalent across the enterprise landscape.
A Single Universal View of Each Customer
The most common and traditional approach to creating a single authoritative system of record involves building from scratch a massive data warehouse containing all of the data that exists in the multiple and often duplicative databases and CRM systems across the enterprise. The problem with this approach is that, first, it would take years to build, map and migrate all of the data from the various sources. Second, as is true of just about every enterprise data warehouse project, the undertaking would carry a multimillion-dollar price tag and suffer delays and cost overruns before eventually being discarded as a failure when (inevitably) turnover in the executive ranks leads to the departure of the original champions and the arrival of a new regime with different ideas.
The other approach — the one I recommend — is to create a data hub that draws from all of the sources and makes all of the data accessible for analysis and use in initiatives such as the delivery of personalized customer experiences. A data hub supports three core functions: the movement, harmonization and indexing of data. Here are examples:
- Data movement involves copying data from each silo and storing it on a new set of servers. Because disk space is so cheap these days, this is an easy solution that eliminates the need to constantly hit the production servers used by each silo across the company.
- Data harmonization is what it sounds like: moving from apples and oranges to apples to apples. It is the ability to harmonize all of the data, regardless of source, and address the naming, structural and semantic differences in data from various sources. Data harmonization allows you to view, use and act on all the data in the same way.
- Indexing enables fast lookup and fast analysis of data in any indexed field. In the old days of relational databases, you would launch a query and then go get a cup of coffee while you waited for the result. With an indexed data hub, all of the data is available in real time. This approach will allow for the introduction of machine learning and artificial intelligence tools that sit on top of the indexed fields.
In the data hub model, siloed data is moved to one place and then it is harmonized and indexed. This is not only the preferred approach, but also the most powerful option and the easiest to build and maintain — all of which translates to a lower total cost of ownership and a higher return on investment. But the most important reason for building a data hub is that it eliminates the siloing of data and enables large enterprises to take unified approaches to their customers.
Think about what that would have meant for the farmer in Iowa: Instead getting 200 emails with confusing and contradictory messages from one company in one week, he would have received a lighter stream of personalized and relevant communications that built brand loyalty. Save the farmer, save the world.