On Information Silos and Fracturing the Enterprise Brain

Those of us who advocate for a move to the cloud often talk of cloud applications ending enterprise silos – the malaise where different data is stored in lots of different applications and where different departments don’t use any pan-organizational system to communicate and collaborate. The theory goes that by utilizing cloud solutions, organizations gain the ability to communicate better and share information more readily.

All of which sounds good, but a recent post over on ReadWrite bought up a valid point, the contention of Huddle CEO Alistair Mitchell that, over the next twelve months, organizational knowledge, will fracture. Mitchell’s reasoning lies in the very real trend for bottom-u adoption of cloud applications, and the corresponding burgeoning of the number of different tools being used across an organization. As he puts it:

When employees are storing stuff in the cloud, and using something like 15 different cloud storage tools to do it, their corporate knowledge, their brain, is destroyed

In the article, Brian Proffitt rightly points out that historically, data has been stored in one of two places – within on-premise servers or on employee’s local computers. So long as the organization had some decent enterprise search or document management in place – organizational knowledge was generally protected and accessible.

Peer inside a large enterprise today and it’s a fair bet that there will be organizational knowledge stored across a plethora of services – cloud storage offerings (DropBox, Google, Box, Huddle etc), point solutions (Project management, document collaboration) and applications of many different types. Proffitt suggests that these disparate systems create a large amount of time wastage:

On a day-to-day basis, workers themselves increasingly struggle with the problems of finding data. Where did I put that latest template for the expense report? My document folder? In the project folder? In Dropbox? Oh well, I’ll just email HR for another copy. Time wasted, again, and redundant effort from the worker and the people that have to help workers replace data they already have. Yes, this might take just five minutes, but how many times does this happen every day?

MyPOV – Bring on the Global Services

One of the recurring themes I’ve been articulating over the past few years is the opportunity for services that span a number of different applications. These can take many different forms – at an infrastructure levels it’s tools like RightScale and enStratus giving organizations a single view across multiple infrastructure assets. For developers, tools like Appsecute (disclosure – I’m an investor) allow them to tie together all the myriad of systems they use into one stream.

At an application level, social tools like Chatter and Yammer allow organizations to pull in content from different teams, locations and applications in order to create a “single pane of glass” at least at a social level. The same trend will rise with documents and institutional knowledge. Already companies like Huddle and Box are trying to solve this dilemma (Huddle through deep technologies which surface relevant content, Box by creating a document collaboration platform which spans an entire organization’s software tools). But discrete tools add to the problem, as Proffitt points out:

But here’s the rub: With the introduction of every new separate social, collaborative or storage service, the problem actually gets worse, because it sets up yet another location for data to live. Without a truly universal service-bridging search tool or the potential to integrate these services, the fracturing of the enterprise brain will only continue

I suspect that, over time, a number of vendors will begin to surface who deal with this information discovery problem without the being conflicted by attempting to also fulfill a specific application space. Both Huddle and Box for example are trying to be content collaboration tools, while at the same time tipping their hat at the cross-application problem. It’s hard to legitimately deliver a pan-application document solution when you’re also trying to be one of the many systems of record. Of course all these vendors are articulating this problem and rapidly telling enterprises that the ideal solution is to deploy their tool across the entire organization, but it feels that, for the time being at least, that isn’t a likely scenario.

Rather solutions that take a Google search appliance approach and apply it specifically to the discovery, archiving and accurate classification of organizational knowledge will be an interesting area of development over the next few years and, in solving that problem, the plethora of data sources and types we need to deal with on a day to day basis.

6 Comments
  • I recall Hasso Plattner saying fractured billing would lead to fundamental changes in business model. Customers seem to be surviving the billing from many differant SaaS-Cloud services. But I can’t see how they can survive for more than a couple of years before a fractured enterprise brain breaks the camels back?

  • Well, I think you hit on one reasonable answer at the end…search. Whoever best solves this problem will do it using search. As for on-premises storage of unstructured data, it is and continues to be a mess. It will probably be easier in the short run to keep all of that unstructured data in a few cloud storage locations and use search to find it. Another way to utilize data that is stored in different cloud apps is to choose choose cloud apps that have an ecosystem built around them. Knowing how your cloud apps can use data in other cloud apps via various APIs or integration points is one way to leverage access to this data no matter where it is. The problem Mr. Proffitt describes is a short-term cloud app growing-pain situation that will very likely yield to solutions and search will be one of the keys.

  • Ben, In large Fortune 500 / Global 200 enterprise organizations, is the problem root more architectural rather than tooling?

    Search/Query and APIs solve discovery and access issues, yet deeper silo fractures exist that are not solved by a search applicance.

    For example:
    1) Semantic faults
    2) Information synchronization / caching disconnects
    3) Data lineage

    Do you believe beyond search, archive, and classify that Enterprise Integration Patterns, Master Data Management, Data Integration, Data Semantics, and Data Service processes are required to unify disparate silos?

  • It is frustrating when I have to go to two different systems to find basic information, I always find myself needing the “mashup” view – I wanted one screen with all the sales information from Salesforce, the customer support tickets from Zendesk, and any emails I’ve exchanged with the customer from Google Apps, ect. A search solution is what I use to solve this.

  • It is frustrating to have to go to two different systems to find basic information, I always find myself needing the “mashup” view –- I wanted one screen with all the sales information from Salesforce, the customer support tickets from Zendesk, and any emails I’ve exchanged with the customer from Google Apps, ect. A search solution is what I use to solve this.

  • It is frustrating to have to go to two different systems to find basic information, I always find myself needing the “mashup” view –- I wanted one screen with all the sales information from Salesforce, the customer support tickets from Zendesk, and any emails I’d exchanged with the customer from Google Apps, ect. A search solution is what I use to solve this.

  • Pingback: Cloud Channel Summit – Do Partnerships Matter More or Less in the Cloud? | SnapLogic Blog

Leave a Reply