Prompted by the recent NYT article about mashups, Anant posted on the difficulty in establishing network effects for data within the Enterprise. He describes the fundamental problem that the dataâ??s all locked up.
I agree 100%.
This is a problem Iâ??ve been thinking a lot about and Iâ??d like to share some simple observations.
There are lessons to be learned from the how network effects took over on the public Internet. If you could identify what caused them to happen there, the chances of being able to replicate them within the Enterprise vastly improve. As I studied this problem something very important occurred to me:
The virtuous cycle of content creation on the Internet that was enabled by interoperability (HTTP and HTML) as well as self-serve discovery (Search/Google), access (Browsers), production (vi/Eclipse/Dreamweaver/Flash/etc.) and publication of content (WordPress/Joomla/Apache/etc.) .
Do any of these exist within the Enterprise?
- Interoperability: Nope. Sure, at the network level thereâ??s interoperability, but each siloed application has its own schema as well as its own API. Data Interoperability along the lines of what HTML did for the public Internet is nowhere to be found for data within the Enterprise.
- Self-service Data Discovery: Nope. A 2007 Accenture study reported that 40% of the respondents said that other parts of the company are unwilling to share information. Managers reported that half of the information they obtain is of no value to them! And most thought that it was easier to get information about their competitors than it was their own organization. To many, if Google hasnâ??t indexed it, it doesnâ??t exist.
- Self-service Data Access: Nope. The great thing about Google being able to find something is that it means that I can access it too. Except for websites, how much data is really accessible within the Enterprise without sophisticated programming skills? Hardly any at all.
- Self-service Production: Here the Enterprise at least has some capabilities in the form of popular Desktop applications like Word and Excel as well as wikis, blogs and other emerging collaboration tools. But without the data thatâ??s locked up in silos, these tools are limited to what users can scrape of websites, cut/paste from other apps or generate by hand.
- Self-Service Publication: Here again the Enterprise has made some progress with blogs and wikis, which might be OK for document publication, but what about the data itself? Is a wiki really a good way to publish data? I donâ??t think so. Wiki data is not readily consumable by other applications. Wikis tend to contain data that previously would have ended up in Word docs. Now that data is easier to find and edit. But wiki-style publication still doesn’t address all that valuable internal data locked away in silos.
So, I agree, data is still locked up, despite the advent of wikis and web sites, and despite the millions of dollars spent on building data warehouses whose data ends up being available only to a few analysts.
I donâ??t think network effects for data within the Enterprise will ever take hold until the barriers to the virtuous cycle I outlined above are removed.