EIP--More Profitable For Integrators Than Users?

Last month, we took a look at the Hummingbird Enterprise Information Portal as an example of one of the more complete offerings in this emerging market, but as a VAR or integrator, knowing about the products that vendors are offering is only half of the profit equation and the lesser half at that. What's more important is knowing what your customers want.

So this month, I'd like to share with you some preliminary results from the Enterprise Information Portal survey now underway at Survey.com. The survey is being solicited from decision-makers in private industry, government, and education and, even though we only have a few hundred responses so far, an interesting picture of the demand side of this market is starting to emerge. What follows is a brief overview of what buyers think about enterprise information portals, which may be useful to integrators interested in positioning themselves to profit from this market. As a bonus, I'll finish up with a little bit of additional information about open-source Unix (which respondents rate among the top three operating systems that an EIP should support) that we've garnered from a brief follow-up to our big survey at the beginning of the year.

Just Looking, Thanks

The first and most important fact about EIPs is that they are still very much an early adopter market: over half of all respondents are merely thinking about an EIP--not even talking to vendors yet. Almost 10% have no intention of implementing an EIP, mainly because they don't see a business case for it or lack the personnel to deal with the issues involved in such a project (an opportunity). Even fewer are presently implementing one. There is apparently a lot of education needed before most people will be ready for this technology.

Among those respondents considering or actually deploying an EIP, the most common scenario for the rollout is a phased deployment by business unit, rather than by division, office, or enterprise wide. Line-of-business managers rank fairly high for their involvement in the development process and in the importance respondents place on their receiving information through a portal. Likewise, business development personnel rank third (behind IT and corporate management) in the vendor selection process. All this underscores the necessity of domain-specific knowledge for EIP integrators and the importance of cultivating contacts with line-of-business managers who are likely to cast the deciding vote for a given solution.

Dispersing The Cloud Of Unknowing

One of the problems afflicting the enterprise information portal market is the lack of a good definition of an EIP. As well, every vendor has a different slant on what an EIP is supposed to do, what kind of information it should deliver, and who should get the information. All this, of course, will be determined in the end by customer demand.

What the survey shows is that, given a choice of several descriptions of an EIP, respondents are most likely to identify it as something that "acts as a single point of access to internal and external information" or "gives users access to disparate enterprise information systems." All well and good, if not terribly useful, except for marketing documents, but we probed many other aspects of EIP operation, as well.

For instance, the survey shows that the most important capability of an EIP solution is its ability to access structured data such as that from a data warehouse or business intelligence application. Respondents also considered the control of user access to information quite important; given this, it's not surprising that they also considered support for LDAP an important factor in choosing an enterprise information portal. The ability to search across all information sources also ranked high.

In the preliminary results, respondents expected that about 60% of the data delivered through their portal would be structured (e.g., database, data mart, etc.), 30% unstructured (e.g., text, email, etc.), and the balance commercial (e.g., news feeds, etc.). However, we expect that the proportion of unstructured data may rise in the final results, although it is likely that structured data will always predominate. Other research we have done indicates that much of the impetus towards EIPs comes from the business intelligence side of companies.

What Have You Done For Me Lately?

Asked what they expected an enterprise information portal to do for their organizations (choosing from a list), respondents said that enabling access to relevant information was their top expectation, followed by faster, better decision making and saving end users time. However, despite the benefits they see in an EIP, almost 50% of them expect it to be a cost center, not a source of profit. On the average, respondents expect an EIP to take between nine and ten months to pay for itself.

Almost a quarter of all respondents either expects to or did outsource the design and implementation of an EIP solution. Combined with the fact that the top three capabilities for an EIP, according to respondents, are application integration, ease of use, and integration with back-end sources, this is good news for integrators in general and it's not likely to be a short-term assignment either. Respondents expect that deploying a 100-user EIP will take about a year and a 1,00-user ETP almost two years.

The three top sources on information that respondents want to make available through an EIP are online access to a data warehouse or data mart, internal documents, and financial information. Workflow was the top collaborative capability required of an ETP, followed by the ability of users to publish information from their desktops.

An enterprise information portal can be used in a variety of ways: for Business-to-Business (B2B) e-commerce, Business-to-Consumer (B2C) e-commerce, and (what used to be called an intranet) Business-to-Employee (B2E) e-commerce. Survey respondents estimated that almost half of the activity of their EIP solution would be focused on B2B functions, a little bit less on B2E, and about a third on B2C. (The total is more than 100% since respondents were not constrained to make things add up to exactly 100%, but rather to furnish ranges.)

Putting It All Together

We also asked respondents a series of questions about the technologies they expected or required an enterprise information portal to support. The top proprietary technologies were Java, "intelligent agents," and, good news for the embattled giant of Redmond, Microsoft Exchange and Microsoft Active Server Pages. The top open-systems technologies are XML (almost 90% demanded this), CORBA, and LDAP. A whopping 88% demanded support for public key encryption, as well.

Microsoft, Oracle, IBM, and SAP were the top vendors whose products that respondents wanted an EIP solution to support. Not surprisingly, Microsoft Windows NT/2000 was the top OS that an ELP should run on, followed by Sun Solaris and open-source Unix (Linux or BSD).

The Open-Source Bonus

That brings me to the OSU bonus I promised. We did a brief follow-up survey to identify the top OSU distributions and how IT decision makers expect or want to buy and get support for open-source Unix. Not surprisingly, Red Hat was far and away the dominant OSU in organizations--the real surprise was the strong showing of Mandrake and FreeBSD in the number two and three positions.

Yet the news is somewhat mixed for VARs and integrators. Overwhelmingly, respondents prefer to buy their Linux or BSD from an open-source Unix vendor--VARs and systems integrators come in at the bottom of the list. The same holds true for how they want or expect to get support. However, it's likely that this reflects the do-it-yourself nature of the OSU market so far--the data hints that, as the market matures, VARs and integrators of all stripes will play an increasingly larger part.

Since we queried purchase and support preferences, not for OSU solutions, but the OS itself, this says little about the actual value add of an integrator selling a vertical solution. As I've pointed out before, there are actually a lot of vertical applications out there that run on Linux--Pick Systems alone offers upwards of 4,000 different ones in a host of different industry sectors. So OSU offers a good way to up profits by subtracting the cost of the OS from the cost of a solution.

Dave Trowbridge is the senior analyst at Survey.com, a market research firm specializing in data-intensive reports on information technologies, where he monitors operating systems, portals, and various aspects of business intelligence.