It was a dark and stormy market; the revenues rose to frothing peaks like the waves of a March nor’easter – only to crash down upon the shores of reality, where vendors consolidated like waves blending into a shimmering, watery shingle before fading into the moonlit sand.
Sorry for that, but it’s easy to be melodramatic about the “frothy” tech market of the late 1990s (or anything for that matter; see the Bulwer Lytton Fiction Contest). I learned a lot about this cycle as the portal analyst at META Group as the portal market rose and dissolved.
The market started with 21 viable, horizontal products, which is too many for any market to sustain. My 2002 METAspectrum (the equivalent to a Gartner Magic Quadrant) had 19 vendors. Then 14 in 2003 and, for the last one in 2005, there were a cool dozen. There was a big convergence and buying frenzy after that. Oracle had four portal products of its own in 2007 after acquiring BEA.
Looking back now, there are several lessons that I learned:
- There is a legitimate need for an aggregation and access point for all the less structured information work. I’m still convinced this underlying concept was sound. Workstream collaboration (such as Slack and Microsoft Teams) seems to be taking up that mantle now.
- When a vendor says “the implementation costs have to be high because this type of product is so customized for your business”, don’t believe it. Implementation costs ran 7x license costs in the earlier days of the market with the explanation that only a tailored fit was worth buying. But sure enough, implementation declined to 3x license costs once the market matured.
- It’s easy to get crazy high ROI results when the process you’re replacing is stupidly inefficient to begin with. It was common to publish case studies or promises of ROI in the hundreds of percentage points. Whenever I dug underneath the surface of those, there was always a prehistoric process that could have been cleaned up any of a dozen ways, and usually one much cheaper than a portal product.
- Diminishing returns on new features hit this market about 3-4 years in. At that point the new features were more likely to be window dressing that went unused. The fancy features were there to demonstrate vitality, create competitive challenges for competitors, provide unique differentiators in presentations and POCs, and goad competitors into spending more on R&D.
- Once products features began to converge – which happened quickly given the level of R&D, investment, and spending – brilliant implementations were more attributable to favorable conditions on the buyer side rather than the product. Case studies of companies doing wonderful, transformative things with a particular portal product were really the result of a combination of a good designer, proper planning, users and a business willing to change, adequate investment (in time, people, and money), well maintained infrastructure, and a visionary program owner. With those elements in place, other products could have yielded the same great results too.
- Consolidation happens within companies as well as markets as products converge and reasons for maintaining multiple products in the same category dissolve. This internal form of consolidation drains the market of revenue needed for R&D and effectively doubles the rate of convergence, moving the market from the growth phase to a more defensive profitability phase.