Is your organization’s data getting lost in translation?

You might know this phenomenon by another name: the dreaded data silo.

Sad to say, most organizations struggle with barriers that impede efficient, transparent data flows. Over time, these barriers harden, becoming all but impenetrable to those on the outside. Every team is an island fortress, jealously guarding the data within.

You know that your organization would be better off without data silos. Without data that gets lost in translation. Here’s how to break down those barriers and ensure they don’t return.

1. Leverage a Platform That Connects All of Your Applications 

The best way to ensure your organization’s data doesn’t get lost in translation is to ensure all its data sources speak the same language. Data silos can’t form when the raw material to construct them is nowhere to be found.

You could hire an army of engineers or devops to manually knit together the dozens or hundreds of apps your organization relies on. But that’s an expensive, dragged-out process. A better solution is to knit them at scale, using a partially automated integration platform as a service, or iPaaS.

An iPaaS, according to Workato, is “a cloud-based platform that allows your organization to build integrations between cloud and on-premise applications.” The platform enables the creation and deployment of robust inter-app data flows that are mutually intelligible. In short, it helps apps that look nothing alike speak the same language.

And in so doing, it has the potential to demolish your organization’s most troublesome data silos.

2. Back Up Your Integration Platform With Applications That Have Built-In Integration Capabilities

A well-utilized iPaaS solution can take you a long way toward a future free of data silos. But unlike a traditional journey or race, the risk of backsliding is real. You need to work to maintain your organization’s hard-won progress.

To do that, you’ll want to select applications that integrate readily with others. Realistically, this means limiting future software investments to cloud-based apps, which are much easier to integrate than legacy systems. There are plenty of other reasons to move away from legacy apps, as you’re well aware.

3. Pare Back the Number of Unnecessary Applications Your Organization Uses

This could be easier in theory than in practice. However, it could be worth investing your clout in what’s likely to be a long, difficult initiative to right-size your application footprint. Because it’s very likely that your organization is “over-apped” right now.

Excessive use of redundant applications has been an issue for years. According to data from F5, the average enterprise organization used something like 122 cloud-based apps in July 2016. It’s virtually certain that figure has grown since then to keep pace with increasingly complex business needs.

To pare back unnecessary applications while avoiding cuts to mission-critical apps, you’ll need to devise a complex set of tests for each app. Don’t lose sight of the big picture, though. Your right-sizing should leave your organization’s app cloud leaner and more cost-effective than before.

4. Limit Use of Unstructured Data (And Work to Structure Existing Unstructured Files)

 The vast majority of the data used by enterprises today is unstructured. It’s not quite fair to say that unstructured data is a “mess” in the informal sense of the word. But it is true that unstructured data is more difficult for organizations to extract value from. That makes it inherently less valuable and less useful than structured data, its more orderly counterpart.

Common types of unstructured data include:

  • Email and workplace chat messages
  • Lists or tables in word processing documents
  • Audio and video files
  • Image files
  • Web pages and HTML files
  • Slideshows and other presentation formats
  • Raw PDFs

That’s just a sampling. Any file format whose data can’t easily be imported into a database is by definition unstructured. As are analog stores of information, such as handwritten notes, paper receipts and invoices, and paper schematics.

Your unstructured data costs you, whether you know it or not. If not directly, then indirectly, through the opportunity cost of its unrealized value. Work to capture that value by methodically converting your unstructured data to structured formats. Overwhelming as that sounds, it’s clearly in your organization’s long-term interest.

5. Use Structured Data Best Practices to Organize Your Data at Scale

 Upgrading your organization’s data to structured status isn’t enough on its own. Your teams need to follow structured data best practices to get the most out of the change. This means:

  • Using consistent formats and conventions for dates, original sources, file names, and other bits of identifying data or metadata
  • Using file versioning conventions, notably renaming the file (“version 3,” “version 4,” and so on) after every change
  • Using descriptive file names that include as much information as possible about the data contained within
  • Creating a universal data dictionary that defines every dataset, file type, and category used by your organization
  • Using hierarchical or relational databases wherever possible, rather than information-poor “flat” databases

6. Assign “Data Liaisons” to Keep Data Silos From Re-Forming

 No matter how good they are at their jobs, database managers and devops aren’t enough to prevent data silos from re-forming. Nor are the automated integration protocols and structured data best practices you should now be using.

Plain and simple, your organization’s data needs a human touch. And who better to provide it than a “data liaison,” a position conceived to ensure accountability for inter-team and -department data flows?

The ideal data liaison is technically proficient and well-versed in data management conventions. But they should also be an effective communicator, someone who can speak the human language of the teams they interface with even as they address their technical needs.

Larger organizations need more than one data liaison to manage their flows. Where data transparency is absolutely critical, one per team might not be too many.

Don’t Let Your Data Get Lost in Translation

Your data is too important to get lost in translation. Fortunately, you now have a blueprint for ensuring that doesn’t happen.

Just remember that data silos are less like buildings than like weeds. They’re simple enough to remove but liable to return without proactive maintenance. If you really want your teams speaking the same language, you’ll need to return to these practices again and again.

Leave a Reply

Your email address will not be published. Required fields are marked *