Business Intelligence Lessons from Softwar: A Portrait of Larry Ellison And Oracle

History has this uncanny tendency to repeat itself.

One of the nice things about operating in a domain as old as business intelligence (“About 60 years!” as our CEO, Vincent, likes to say) is that you can look back at the long arc of history and figure out what’s been tried and what hasn’t by the vendors of decades past.

Softwar: A Portrait of Larry Ellison and Oracle is one such blast from the past. It was written by Matthew Symond (with footnotes contributed by Ellison), and was published in 2003. The book contains a look at Oracle through the events of the 2000 dotcom bubble, the aftermath of the 9/11 terrorist attacks, and Ellison’s attempt to have Oracle break out from databases and into the enterprise software market with its Oracle E-Business Suite.

And boy does it have useful nuggets for those of us in business intelligence.

Silos are Old, and Ellison Hated Them

When Softwar was written, the enterprise applications business was mostly dominated by best-of-breed systems integrators like IBM, PriceWaterhouseCoopers, and Anderson Consulting. These integrators would select a collection of enterprise software, and then customize it for each customer.

Ellison hated this. He thought such custom work was extremely wasteful, not to mention extraordinarily expensive. Of course, Oracle had a systems integration services arm themselves, but they weren’t anywhere near the largest integrators in size, so Ellison was motivated to try and change the rules of the game. But his dislike came from a kernel of truth: most integration projects did take multiple years, and cost its customers hundreds of millions of dollars. It gave software a bad name.

Ellison’s take was that people should shift to all-in-one software suites, where the suite provider could integrate everything into one consistent data model. This made it cheaper on enterprise customers:

The systems integration industry has successfully taught enterprise software customers how to buy software so that costly customization is an inevitable outcome. Typically, prospective users put together a team to draw up a Request for Proposal (RFP) of desired features and functionality. As a rule, partly these RFPs reflect the way the firm has traditionally done business, and partly they are a wish list of features that often quite low-level people suppose will improve the operation of their small bit of the organization. The result is software that is expensively adapted to lock in outdated and ill-considered business practices—a major reason why ERP installations have only rarely given an adequate return on investment.

With the E-Business Suite, Ellison wanted to turn that practice on its head. Instead, Ellison wanted customers to change their processes to take maximum advantage of the Internet-based best practice captured by the software. Ellison said, “Don’t tell us how you’ve been running your business for the last twenty years. Instead, let’s try to figure out how you want to run your business for the next twenty years. It’s a classic business mistake to say, ‘This is how we do business; change your software so we can automate it.’ A new approach is needed. First, you must simplify and modernize all your business processes, then move those newly standardized processes to the Internet. Only then can you expect the system to improve your business. Only then can you expect the system to be delivered on time.

Of course, such cost savings weren’t the only thing that Ellison was betting on. Halfway through the book, Ellison realised that it was only possible to sell his vision of a unified enterprise suite if he could offer a compelling benefit on top of that. He found one with business intelligence. Symond writes:

Mark Barrenechea, a chubby, confident young programming manager who had come to Oracle in 1997 to build a CRM development team, remembers the moment when the truth about the applications business first hit Ellison. “He asked a very good business question. He said, ‘How many employees at Oracle do we have?’ Because, logically, as in all complex networks, everything is centralized. He just assumed HR data was centralized. So he asked how many employees there were at Oracle. Ron replied, ‘Well, we’d have to put a data warehouse together. We’d have to go out and put a team together. We’d have to consolidate formats. Larry, in ninety days I’d give you an answer’ and the guy went off the wall. He said, ‘Why isn’t all this data in one place? We have the Internet now, right. We can consolidate this into one global schema.’

Ellison began to push his team to build business intelligence capabilities on top of a standardized data model, across every business flow they had built into the E-Business Suite. In January 2002, a few years after this initial realization, Ellison was ready to unveil what they had built:

The Daily Business Intelligence, which Jarvis and Ellison were demonstrating in Amsterdam, was a dramatic illustration of what became possible with a single data model. It was one thing to show slides illustrating how SAP’s applications ran across five separate databases linked by clunky messaging compared with the Oracle suite’s tight integration with a single database. It was rather more impressive to be able to offer something that was both spectacularly useful and beyond SAP’s capabilities because of the inferiority of its information architecture. The same could also be said of business flows, but Daily Business Intelligence was a much easier concept for most people to grasp and thus to sell to them.

(…) Ellison used his keynote to explain why the Daily Business Intelligence was not just another nice feature but was pretty close to being the Holy Grail of business IT. As usual, he blamed most of the problems afflicting business computing on the fatal combination of fragmented information and incomplete automation: “Most business processes are not completely automated; they’re a combination of online and offline activity. For example, in the sales process called ‘opportunity to order’ each line of the sales order is always entered into the application system, but the price quote and sales contract might be stored separately in an Excel file or a document image file. If you store your price quotes in Excel files, that means your applications database is incomplete, so you can’t ask the system questions like ‘What is the total value of all my outstanding quotes?’ Similarly, if you use a document image file to store contracts, you cannot ask the system, ‘How many of my contracts have nonstandard limitation-of-liability clauses?,’ because that information is not stored in the applications database. In other words, if your process automation system is incomplete—and most are—then your applications database will be incomplete as a consequence. And an incomplete database makes building a business intelligence system virtually impossible.”

It’s the “completeness” of the E-Business Suite’s process automation and the single-data model, which allows you to keep all your information in one database, that makes Daily Business Intelligence possible. “Every day, Oracle sales managers see updated sales figures and sales forecasts. Every day, salespeople see how much they sold and how much they spent so far this month, this quarter, this year. They see exactly where they rank among their peers in sales and expenses. If they’re in the lower ten percent in sales and the upper ten percent in expenses, that fact greets them every day on their home page. If you’re an Oracle engineering manager, you see how well your product is selling, and you see how many service requests your product is generating. If your product is in the bottom ten percent in sales and the top ten percent in service requests, you know you have a serious quality problem that needs to be fixed. The Daily Business Intelligence system not only provides managers with up-to-date data to make better decisions, it provides every individual in Oracle with better information—so they know what to do and how well they’re doing it.”

As well as providing detailed, almost instantaneous information about every aspect of Oracle’s business, it also has a dramatic impact on the behavior of individuals: “Perhaps the most interesting aspect of Daily Business Intelligence is the comparison of your job performance with that of your peers. The reaction to these rankings is fascinating. People not only work harder to move up in the rankings, they work smarter. They study the people with the best performance and learn from them. And everyone gets daily feedback that lets them know if their job performance is improving. This makes a manager’s job much easier. The manager doesn’t have to constantly tell people that they’re doing a good job or a bad job; the system does that. When Sally arrives at work and logs on to the system, her home page pops us and says: ‘Hi, Sally, you rank in the top twenty-five percent in sales, congratulations, but you’re in the top five percent in entertainment expenses—knock it off; stop eating at all those fancy French restaurants.’ But you know what, the second that Sally realizes we can see how much she’s spending on expensive dinners, her behavior changes . . . she stops ordering the expensive wine with dinner. Constant feedback, whether it’s from a boss, a parent, a teacher, or the E-Business Suite’s Daily Business Intelligence system, encourages people to improve their behavior. Hard facts about how well you’re doing your job in comparison with other people in your company takes most of the human bias out of performance assessment. It doesn’t matter much if your boss likes you or dislikes you, it’s all about your numbers.”

What Happened To Ellison’s Vision?

Of course, we know today that Ellison’s vision didn’t play out the way he thought it would.

In 2002, Ellison firmly believed that the enterprise software space would consolidate into a handful of providers — Oracle and SAP amongst them. Everyone else — Peoplesoft and Siebel and i2 — would be bought up by the giants. (Indeed, two years after the book was published, Oracle had acquired Peoplesoft and Siebel for a discounted price, reflecting their failure to compete).

But what Ellison didn’t see was the rise of SaaS. He mentioned ‘software as a service’ only a handful of times in the book, and by 2003 had already been an early investor in former Oracle exec Marc Benioff’s Salesforce. (Salesforce IPO’ed a year later, in 2004). But Ellison didn’t see that the rise of the internet would expand the enterprise software market, not consolidate it! Salesforce’s rise, — and Atlassian and Workday’s and Slack’s rise in the years afterwards — prevented Oracle from shifting the market to all-in-one software suites, with standardised data models.

In this case, however, the shift to SaaS made perfect sense: SaaS was simply a superior business model for enterprise customers, since it reduced the up-front costs of software. With the benefit of hindsight, the rise of SaaS was as inevitable as it was unstoppable.

As a result, today’s data world looks very much like the past. Sure, we don’t have as much data locked up in expensive on-premise software suites, running on locally installed databases, kludged together with custom code from integrators. But we haven’t moved that far away from that reality either. Today’s SaaS-heavy software stack has just as many data fragmentation problems as the stacks of the past.

The best hope we have is the shift to API-first SaaS companies. In a recent Invest Like The Best interview, Benchmark Capital’s Eric Vishria argued that the next generation of enterprise SaaS companies they were investing in were starting to take API-first product development seriously. Perhaps this is a way out for our data fragmentation problems — you don’t have to worry about silos as much when there’s an easy way to consolidate data in a data warehouse.

Wrapping Up

In August this year, Oracle announced the end-of-life for Daily Business Intelligence. It wasn’t the holy grail they were hoping for. The enterprise software industry had changed, and Oracle had to change along with it. Larry Ellison himself stepped down as CEO of Oracle in 2014, after a decade more of high-stakes competition.

Oracle, by all accounts, remains a hugely successful company. But it’s telling that while battlegrounds may shift, many of the playbooks appear the same. In fact, one of the most interesting nuggets in Softwar are mentions of people who continue to play an outsized role in enterprise software. In Chapter 27, when Ellison is asked about potential successors, he tells Symond:

When pressed, Ellison was prepared to talk about only two of his managers who might turn out to have the necessary qualities. One was Thomas Kurian, an Indian in his mid-thirties who had arrived at Oracle in 1996 via Princeton, McKinsey, and Stanford. Kurian was intense and serious—a little dour, even, although Ellison assured me that he’d managed to make him laugh at least once. But Kurian, in only a few years, had achieved something remarkable and of incalculable strategic importance to Oracle: after its successive failures in building a competitive application server and the departure of those who had been responsible, Kurian had made Oracle, according to its own numbers, second only to the incumbent BEA in the application server market. As well as being a fast-growing and, in time, juicy new revenue earner, the application server was critical to Oracle’s ambition to be the dominant enterprise software infrastructure provider committed to using open standards. Kurian had created a strong engineering team, driven it with a combination of discipline and pragmatism, produced a quality product, and then sold it with passion and energy to the outside world. When Ellison mentions Kurian, he can’t help beaming “Thomas has done a brilliant job, absolutely extraordinary,” he says.

Today, Thomas Kurian is the CEO of Google Cloud. He was brought in to Google to contest a three-way fight between AWS, Microsoft, and GCP in the cloud infrastructure space.

Two decades have passed, but some battles, and some fighters, remain the same. Softwar is a fascinating book, and the fights that it chronicles should feel familiar to many of us in data. I can’t recommend it highly enough.