It’s oddly difficult to find a detailed history of Business Objects, a seminal forerunner to the modern wave of self-service data analytics. The company was acquired by SAP in 2007, and most people nowadays only know about Business Objects as part of SAP’s “BODS” package - unaware of its standalone contributions to the software industry. Bernard Liautaud, co-founder and former CEO of Business Objects, gave a very insightful talk at the London Business School in 2014; from the recording, I was finally able to piece together the story.

Rewinding to 1990: Oracle was an emerging software juggernaut, spreading the gospel of the relational database throughout the largest public and private organizations in the world. SAP and other transactional software packages were successfully building business applications atop the new database technology. The world had collectively awoken to the fact that they needed robust, digital appliances to store their most important information. Bernard Liautaud was working at Oracle in Paris as a pre-sales engineer at this time, observing the mania firsthand.

The explosion in relational database adoption had introduced new ways of working: business users, particularly in areas like finance and logistics, were now able to crunch granular data pertaining to sales and inventory, create high-fidelity reports on the state of the business, and generate data-driven forecasts. The bottleneck, though, was typically the database administrator (DBA); this person managed the enterprise’s databases, triaged the glut of incoming data requests, and then figured out how to sensibly sequence the delivery. Each corporation’s Oracle installation was a production system, often serving as the source of truth for sales, inventory, and other critical functions; irresponsibly running expensive operations against this system could introduce major disruptions to the business - and DBAs served as the first line of defense.

Bernard observed that companies were desperately looking for ways to alleviate this bottleneck. At the time, relief typically only came through hiring more DBAs. SQL was still an emerging standard, and training (and trusting) business users to write well-crafted database queries was a bridge too far, for most enterprises. Bernard happened to meet an independent engineer, Jean-Michel Cambot, who had a crazy idea for enabling the masses of non-technical users. Cambot envisioned a semantic layer consisting of “business objects”, that provided an intuitive abstractions atop the relational database. Using such a layer, business users would be able to craft queries and assemble reports, using point-and-click interfaces rather than code.

After Oracle’s marketing department passed on the idea, Bernard decided to form a new company to independently pursue it. Cambot wasn’t interested in a rigid working structure, but agreed to a 25% royalty on all sales in exchange for building the software - a deal that turned out to be very lucrative for him.

Now on his own, Bernard pitched the initial version of Business Objects to his former coworkers; while Marketing had panned the idea, Sales was willing to listen. Oracle was in a fierce battle with Sybase, and they were hunting for any edge that would allow them to corner the French market. Sales leadership found the pitch compelling; Business Objects would provide a novel “self-serve” analytical capability that extended the core value proposition of the relational database - and critically, for the time being, it only worked with Oracle. This clear alignment produced a half dozen quick sales for Bernard’s fledgling team.

As sales begin to flow, Bernard needed additional funding to properly build out the initial team. In 1991, Business Objects became the first French software company (as far as they knew) to get seed funding ($1MM) from Silicon Valley investors. With the money, they were able to hire ~eight people in France, and critically, prove that they could sell to the American market. $1MM in sales in ‘91 gave way to $5MM in ‘92, and $14MM in ‘93. The value proposition was extremely legible, and their complementary relationship with Oracle continued to provide strong avenues for growth.

In 1994, after four years of growing at 150-200% YoY, Business Objects went public. Bernard and the board elected to debut on the NASDAQ, rather than a smaller European exchange. The drumbeat of being a public company then followed; product lines expanded, along with geographic presence. Bernard recalled that it was fairly smooth sailing for the first 18 months; growth continued at a steady pace. By the end of 1995, the market cap had grown 10X.

Around this time, Bernard and team felt a growing need to revamp the core product. Competitors from North America were gaining serious momentum, and Business Objects’s core analytics and reporting capabilities were becoming less differentiated. The company reorganized around a massive internal R&D project, which they intended to completely replace the existing product line. They were specifically betting on the launch of Windows 95, which promised to bring sophisticated GUI-based computing and a host of new features to desktops around the world.

Despite repeated warning signs from the development team, Bernard pushed hard. The revamp ended up launching in very buggy fashion, and to a much smaller Windows 95 install base than Business Objects had anticipated. Bernard faced anger on two fronts: early adopters of the revamp were understandably frustrated with the quality of the initial release, and legacy customers were growing increasingly frustrated as well. Business Objects had largely neglected the existing install base, as they’d surged to release the revamp. And to make matters worse: as 1996 unfolded, a massive deal fell through in Germany, causing the company to restate their earnings.

The vultures began circling, floating offers to buy Business Objects for ~$100MM - a price tag that would’ve been insulting the prior year. Bernard, feeling more heat than ever, decided to resist the siren call, and instead own up and dig in. The company took a series of dramatic actions: the management team moved to the US to get closer to its primary software partners and investors; the development team axed most new feature work, and focused exclusively on shoring up the quality of the revamp. Not everyone was on board: sales representatives were frustrated, knowing the slog meant they wouldn’t meet quotas. Bernard recalls that things stayed painful for a while, but they were doing what was necessary.

By the end of 1997, Business Objects had managed to come up for air. They’d made significant strides in fixing the revamp (now firmly marketed as their core offering), which had repaired trust both internally and externally. A focused new feature set around making Business Objects reports available on the web, launched first-to-market; the influx of popularity boosted growth by 50% in 1998. By 2000, the company had grown to $300MM in annual sales, with a defensible product mix and financial standing that allowed them to weather the bursting of the internet bubble. As Business Objects continued to grow, bringing in $500M in 2003, the market was undergoing another shift.

The business reporting market was evolving into “business intelligence”, and demand was exploding for interactive web-based analytics, which provided a more dynamic experience than traditional reports. With the Canadian firm Cognos nipping at their heels, Business Objects acquired Crystal Decisions - a pioneer in the realm of pixel-precise, interactive web dashboards. Paired with Business Objects’s core object layer and baseline reporting functionality, the acquisition provided immense leverage. By the end of 2004, Business Objects was in a clear leadership position; and by the end of 2007, they had ridden the veritable hockey stick to $1.5 billion in revenue.

Around this time, SAP came knocking - asking to have “strategic talks” (i.e., probing for an acquisition). Oracle had sent a lowball offer the year before, and Business Objects knew they were in a position of strength. SAP was thoroughly dominating the ERP space at the time, and saw clear opportunity for expansion through tightly integrated business intelligence offerings. They made an offer for $7 billion, a 40% premium on Business Objects’s market cap. Bernard recalls the acquisition was friendly, and made good sense to all parties.

Today, Business Objects remains a pillar in SAP’s core offering. If you browse around YouTube, you can see some of the core concepts still very much alive in today’s product: the concept of data “universes”, that contain data foundations, which in turn house business layers that are full of intuitive objects; Crystal Reports are a click away. Obviously a lot else has changed; cloud-native offerings now dominate the business intelligence market, and new data-driven software services are emerging daily.

But the underlying motif isn’t so different from what Bernard saw almost 30 years ago: the need to abstract away technical complexity, and plug business users directly into an intuitive representation of the underlying data.

Thanks for reading; you can check out my Twitter here