It’s the nature of the tech industry to be consumed with the next big thing. Silicon Valley is particularly guilty of an unhealthy obsession with predictions; guessing the next breakthrough trend, technology, or company before it goes mainstream is a coveted badge of honor.
Never have I felt this anticipation more than recently, as we edge towards the precipice of a big data analytics breakthrough. But as it turns out, if you have a good history of the recent past, the next big trend might not be as hard to guess as you might think.
The buzz around analytics has been going for a while, but there’s no silver bullet yet for any given business or environment. The promise of competitive advantage gleaned from analysis has driven many organizations batty with getting the latest and greatest tools. Marketing babble ricochets across the industry, with each claim louder than the next. Business units across the enterprise are becoming increasingly excited about the prospects that data analytics can offer. But analysis is often a messy process these days.
With technical advances, we’re often too absorbed with what new gadgets are going to produce rather than the more necessary consideration of how things can be produced most effectively. As I thought about this, I couldn’t help but think of how closely related the analytics discussion is to another data analysis advancement of the last 10 years: electronic discovery.
Different song, same tune.
Let’s quickly take a look at 3 of the issues with the evolution of eDiscovery, and it’s parallel to the analytics market. These problems are happening again with analytics because of our rush to play with the new toys instead of stepping back and truly evaluating best practices. The madness is exciting, but don’t neglect the method.
(1) Sandbox vs. Beach
There are great tools on the market that offer analysis for specific legal tasks. Solutions like Clearwell and Recommind provide organizations with the ability to process and analyze content from a data set pulled from an archiving tool, ECM, or other storage platform. The technical ability of these tools is highly regarded, but they can only work with the information they’ve been provided. The heavy lifting is left to the searches that must happen before the data moves to the analysis tools, and this is where eDiscovery often falls short. The native search engines of email and file servers don’t provide the granularity or accuracy. Once the data enters a system like Clearwell, some fantastic analysis can happen. But if you’re putting garbage in, that’s what you’ll be taking out.
There are currently many analytics platforms offering companies incredible insight into data, but none of them can offer the power from an archiving and search perspective necessary to be the starting point of their analytics process. Tableau, for example, provides a stunning ability to render analytics and visualization on multiple data sets, but it needs to be fed that data manually. Like leading eDiscovery providers, institutions should unify their analytics directly into their archiving, providing a holistic view of the insights gained from analytics.
(2) Point Solutions vs. a Centralized System
Anyone familiar with eDiscovery knows of the Electronic Discovery Reference Model (EDRM). It outlines the steps necessary for proper eDiscovery: from Information Governance through Production (additional information here). There are multiple steps, and companies have often employed separate solutions for individual steps, sort of like checking boxes off a list. It works in theory, but neglects that the steps are highly iterative rather than linear. A fragmented system requires data to continually be moved from solution to solution. Anyone with a tech background knows that asking completely different solutions to work together is a recipe for trouble: especially when mistakes have inherent legal repercussions. Soon enough, businesses were clamoring for “end-to-end” eDiscovery tools that claimed to provide all steps under the same roof.
The current analytics landscape is strikingly similar. With the analytics space still nascent, companies are rushing to market with tools that are limited in scope. Log analytics tools and email analytics tools both claim to offer vast insight, but the analysis is limited to specific data. We see the same fragmentations between analytics tools designed specifically for structured and unstructured content, as well. But increasingly, there’s a demand for analytics platforms that can integrate a greater variety of content in one environment, without data movement. A system like ZL UA allows for structured content – such as log data -- to be viewed on the same platform that connects email, files, Salesforce, and IM.
(3) Whack-A-Mole Management
The biggest problem of a fragmented data management system of any sort is the effort required to care for and feed it. Different systems working together require different system “owners” to work together. If one part breaks down, it affects everyone. The analytics systems in place today are no different. Fragmented analytics environments still rely on specialists for each piece. Bringing everything together is ideal. In a perfect world, analytics tools should be brought TO the data, rather than data being moved TO the tools. This allows for better control of data for more insightful analysis… but also reduces the sweat, tears, and time wasted for all parties and specialists involved.
So as you reflect now on your own current analytics environment, maybe you’ll start to feel a sense of déjà vu. After all, we’ve seen this before. And as you probably learned in school, the reason we learn history is so that we may not make the same mistakes of the past.