The Most Important Software Company You Take for Granted
Somewhere around 77% of the world's transaction revenue touches an SAP system. Not an estimate loosely tossed around by analysts — a figure the company itself cites, and one that nobody in enterprise software seriously disputes. Every time you buy a can of Coca-Cola, fill a prescription at Walgreens, drive a car whose supply chain was orchestrated in Wolfsburg or Stuttgart or Detroit, book a flight on Lufthansa, or receive a paycheck from any of roughly 300,000 organizations worldwide, you are — at some remove — transacting through software built by a company that most consumers have never heard of and that even many technologists treat as a punchline about legacy enterprise bloat. The paradox of SAP SE is that it is simultaneously the most systemically important software company on Earth and the one whose brand carries the least cultural cachet outside the corporate back office. It is infrastructure in the most literal sense: invisible, load-bearing, and catastrophically expensive to replace.
That invisibility is, in its own way, the moat. SAP does not compete for attention. It competes for the general ledger. And once it wins, it almost never loses.
By the Numbers
The SAP Machine
€31.2BTotal revenue, FY2024
€17.1BCloud revenue, FY2024
+27%Cloud revenue growth YoY (FY2024)
~437,000Customers in 190+ countries
~107,000Employees worldwide
€280B+Approximate market capitalization (early 2025)
77%Share of global transaction revenue touching SAP
~25Of the top 100 most valuable companies running SAP
The story of SAP is not, at root, a Silicon Valley story — and this matters more than any feature comparison or product roadmap. It is a story about German engineering applied to information, about the conviction that the messiest, most unglamorous problems inside corporations — materials planning, financial consolidation, human capital allocation, supply chain logistics — are also the most valuable to solve, and that solving them in real time, in an integrated system, creates a kind of gravitational lock that no competitor can easily disrupt. It is a story, ultimately, about what happens when five IBM engineers in Mannheim decide that the mainframe batch-processing paradigm is wrong, and that the future belongs to software that processes business transactions as they happen.
Five Engineers and a Conviction
In 1972, Walldorf was a town of roughly 14,000 people in Baden-Württemberg, the southwestern German state that also produced Mercedes-Benz, Porsche, and Bosch — a region whose industrial culture treated engineering not as a profession but as a worldview. Five former IBM Deutschland employees — Dietmar Hopp, Hasso Plattner, Claus Wellenreuther, Klaus Tschira, and Hans-Werner Hector — left Big Blue to found Systemanalyse Programmentwicklung, a name so aggressively literal it could only have been German. They shortened it to SAP.
Hopp, the eldest and the organizational center of gravity, had spent years inside IBM watching clients struggle with batch-processing systems that computed results overnight — meaning that a manufacturer who wanted to know its inventory position at 2 PM would have to wait until the following morning for an answer derived from data that was already a day old. The conviction that animated SAP's founding was elementary but radical: business software should process data in real time, integrating financial accounting, materials management, and operations into a single system that reflected the actual state of the enterprise at any given moment. They called their first product, with characteristic German directness, R/1 — "R" for Realtime, "1" for the first generation running on mainframes.
Plattner, the most technically brilliant of the five and the one whose influence would endure the longest, brought an almost physical restlessness to the project. Born in Berlin in 1944, he had the relentless intensity of someone who experienced postwar reconstruction not as history but as childhood. His instinct was always to rebuild from scratch when the architecture had gone wrong — a trait that would surface, decades later, in the most consequential and divisive bet SAP ever made.
The founding team's first client was the Walldorf branch of Imperial Chemical Industries. They wrote code on weekends, on IBM hardware they accessed after hours — a bootstrapping arrangement that was less Silicon Valley garage mythology and more German pragmatism about resource constraints. By the end of 1972, they had a payroll accounting system running in real time. By the end of the decade, they had something far more important: a modular enterprise resource planning system — though the term "ERP" wouldn't exist for another fifteen years — that could be configured for different industries without rewriting the codebase.
The Architecture of Lock-In
What SAP built between the mid-1970s and the early 1990s was not merely a product but an operating system for the modern corporation. The key architectural insight was integration. Before SAP, a typical large enterprise ran separate software systems for finance, manufacturing, procurement, human resources, and sales — systems that did not talk to each other, that stored data in incompatible formats, and that required armies of staff to reconcile manually. SAP's R/2 (released in 1979 for mainframes) and then R/3 (released in 1992 for client-server architecture) offered a single integrated database and a modular application suite that shared a common data model.
This sounds pedestrian. It was revolutionary.
The integrated data model meant that when a sales order was entered, the system simultaneously updated inventory, triggered procurement if stock was low, posted the revenue recognition entry, calculated the tax implications, and adjusted the production schedule. One transaction, reverberating across every function. For a multinational manufacturer with dozens of plants, thousands of SKUs, and operations in fifty countries — each with different tax codes, labor regulations, and reporting requirements — this was not a convenience. It was the difference between managing by spreadsheet and managing by system.
But integration came at a cost the customers wouldn't fully reckon with until later. Implementing SAP was not like installing software; it was like rewiring the nervous system of the organization. A typical R/3 implementation in the 1990s took eighteen months to three years, cost tens of millions of dollars (often hundreds of millions for the largest enterprises), and required the client to reengineer its business processes to fit SAP's built-in logic — or to customize the software at enormous expense. The consulting firms — Accenture, Deloitte, IBM Global Services, Capgemini — built entire practice areas around SAP implementations, and their fees routinely dwarfed SAP's own license revenue by ratios of 3:1 or 5:1.
Nobody ever got fired for buying SAP. But plenty of people got fired for implementing it badly.
The implementation pain was, perversely, the source of the lock-in. Once an enterprise had spent $200 million and three years configuring SAP to run its global operations, the switching cost was not the software license — it was the organizational trauma of doing it all over again. The data was in SAP. The processes were designed around SAP. The staff had been trained on SAP. The consultants knew SAP. To rip it out would mean years of parallel operations, hundreds of millions more in spending, and a non-trivial risk of operational catastrophe during the transition. Companies that installed R/3 in the 1990s are, in many cases, still running SAP today — upgraded, extended, patched, and migrated, but fundamentally still on the platform.
This is the deepest moat in enterprise software. Not technology. Not brand. Not network effects in the consumer sense. Process entrenchment. The software becomes the company's institutional memory, its encoded logic for how things work. Replacing it means replacing that logic, and no CIO wants to be the person who broke the general ledger.
The R/3 Supercycle and the Making of a European Giant
The release of R/3 in 1992 was the inflection that transformed SAP from a large German software company into a global juggernaut. R/3's client-server architecture liberated SAP from the mainframe, opening the platform to the explosion of Unix and Windows NT servers that defined the 1990s IT build-out. Timing mattered enormously: the early-to-mid-1990s were the Y2K era, when enterprises worldwide were forced to audit and often replace legacy systems that couldn't handle the date rollover. SAP became the default answer.
Between 1992 and 2000, SAP's revenue grew from approximately DM 3.3 billion to over €6.2 billion. The company went public on the Frankfurt Stock Exchange in 1988 and listed its ADRs on the New York Stock Exchange in 1998. By the late 1990s, SAP had installed its software in the majority of the Fortune 500 and had become the third-largest software company in the world, behind Microsoft and Oracle.
Key milestones in SAP's rise to ERP dominance
1992R/3 launched on client-server architecture, breaking SAP's mainframe dependency.
1995SAP crosses 10,000 customers worldwide.
1996R/3 becomes the standard ERP system for the Fortune 500. Revenue approaches DM 5 billion.
1998NYSE listing; SAP market cap briefly exceeds $40 billion.
1999Y2K remediation drives a final surge of on-premise license deals. SAP commands ~30% of the global ERP market.
2000Dot-com bust stalls new license growth; SAP begins tentatively exploring internet-based delivery.
The R/3 supercycle established SAP's business model in a form that would persist for two decades: large upfront license fees, annual maintenance contracts worth roughly 22% of the license fee (providing a predictable recurring revenue stream), and an ecosystem of implementation partners whose own revenues depended on SAP's continued dominance. The maintenance stream alone was a marvel of capital efficiency — high-margin revenue that flowed in annually with minimal incremental cost, essentially a tax on the installed base for the privilege of continued support and bug fixes. By the mid-2000s, SAP's maintenance revenue exceeded its new license revenue, and the company's operating margins in its most mature segments routinely exceeded 30%.
But the R/3 model carried the seeds of its own disruption. The massive implementation costs, the rigidity of customized systems, and the periodic trauma of version upgrades created a latent frustration in the customer base that a generation of cloud-native competitors would eventually learn to exploit.
The Oracle Wars
No account of SAP can be written without understanding its relationship with Oracle — a rivalry that shaped enterprise software for thirty years and that, in its intensity, its personal venom, and its strategic chess, is the closest analog in technology to Coca-Cola versus Pepsi, except that the combatants sold databases and ERP modules instead of sugar water.
Larry Ellison, Oracle's cofounder and chief provocateur, made SAP-bashing a performance art. At Oracle OpenWorld conferences, he would display SAP's product names on screen and mock them. He poached SAP executives. In 2005, Oracle hired away SAP's head of U.S. sales. But the rivalry's most explosive chapter came not from talent raids but from an acquisition: in 2004, Oracle launched a hostile bid for PeopleSoft, one of SAP's most significant competitors in HR and financial software. After a bruising fight — PeopleSoft's CEO literally took out newspaper ads urging shareholders to resist — Oracle prevailed, paying $10.3 billion. In one stroke, Ellison eliminated a major SAP competitor and absorbed its customer base.
SAP's response was a period of frantic M&A. The company acquired BusinessObjects for €4.8 billion in 2007, bringing in the business intelligence market leader. It bought Sybase for $5.8 billion in 2010, gaining mobile enterprise technology and a database platform. And it would later make even larger cloud acquisitions — SuccessFactors for $3.4 billion in 2011, Ariba for $4.3 billion in 2012, Concur for $8.3 billion in 2014 — each deal a response to a different dimension of the cloud threat.
But the Oracle rivalry's most damaging episode for SAP was the TomorrowNow scandal. A subsidiary called TomorrowNow, which SAP had acquired to provide third-party support for Oracle products (a market Oracle considered its exclusive domain), was found to have illegally downloaded millions of files from Oracle's support website. Oracle sued. The resulting litigation dragged on for years, and in 2010, a jury awarded Oracle $1.3 billion in damages — later reduced on appeal, then settled in 2014 for a reported $356.7 million. The case was a reputational stain that felt alien to SAP's self-image as a principled German engineering firm.
The deeper lesson of the Oracle wars was strategic: both companies realized, at different speeds, that the on-premise license model was a melting ice cube. Oracle's response was to acquire its way into cloud infrastructure (buying Sun Microsystems for $7.4 billion in 2009) and launch Oracle Cloud. SAP's response was slower, more agonized, and ultimately more transformative — but it required a specific individual to return to the cockpit.
The Plattner Doctrine
Hasso Plattner had stepped back from SAP's day-to-day management in 2003, taking the title of chairman of the supervisory board. He was, by then, immensely wealthy — a sailing enthusiast who won the 2007 Transpac race, a philanthropist who endowed the Hasso Plattner Institute in Potsdam, a design thinker who funded Stanford's d.school. He could have stayed on the board as an emeritus presence, rubber-stamping strategy while racing yachts.
Instead, he blew up the product architecture.
In 2010, Plattner announced HANA — High-Performance Analytic Appliance — a revolutionary in-memory database that stored and processed data entirely in RAM rather than on disk. The technical proposition was stark: traditional databases, including Oracle's, were built around disk-based storage with complex indexing schemes to compensate for the physical slowness of reading from spinning platters. HANA eliminated that bottleneck. For SAP's ERP workloads — which involved enormous volumes of transactional data that also needed to be analyzed in real time — the performance gains were staggering. Queries that took hours on traditional databases could run in seconds.
We are not just building a faster database. We are making it possible for enterprises to run their entire business in memory. That changes what questions you can ask.
HANA was not merely a database product. It was the foundation for a complete reconceptualization of SAP's application suite — a ground-up rebuild of the ERP system, eventually branded S/4HANA, that could only run on the HANA database. This was, strategically, an extraordinary gamble. SAP was telling its 300,000-plus customers that the future required them to migrate off the R/3-derived systems they had spent decades and hundreds of billions of dollars collectively implementing, onto an entirely new platform. The carrot was genuine — real-time analytics, simplified data models, a modern user experience — but the stick was the eventual end of support for the old systems.
The bet was also a direct assault on Oracle. R/3 and its successors had been database-agnostic; customers could (and did) run SAP on Oracle, IBM DB2, Microsoft SQL Server, or SAP's own MaxDB. By requiring S/4HANA to run on the HANA database, SAP was pulling the rug out from under Oracle's single most profitable customer relationship: the database licenses sold to SAP's own installed base. Plattner understood this perfectly. It was war by architecture.
The Cloud Migration: Controlled Self-Disruption
The transition from on-premise perpetual licensing to cloud subscription — the great migration that reshaped Adobe, Microsoft, Autodesk, and essentially every legacy software company in the 2010s — was existentially fraught for SAP in ways that it was not for most peers. The reason was the maintenance stream.
SAP's on-premise model generated two revenue flows: license fees (paid once, at purchase) and maintenance fees (paid annually, typically 22% of the license fee, for as long as the customer ran the software). By the early 2010s, the maintenance stream was SAP's single largest and most profitable revenue category — roughly €10 billion annually, at margins well above 80%. Every customer that migrated to the cloud would, in theory, stop paying maintenance and start paying a cloud subscription instead. The cloud subscription might eventually be larger in absolute terms, but the transition period would be brutal: maintenance revenue would decline before cloud revenue could fully replace it, creating a valley of death in the P&L.
This is the innovator's dilemma in its purest form, and SAP navigated it with a mixture of strategic boldness and tactical caution. The boldness: CEO Bill McDermott, who led SAP from 2010 to 2019 — a Long Island-born salesman who had run a deli at age sixteen, the first American to lead Europe's largest technology company, a figure of relentless optimism and enormous personal charisma — spent more than $30 billion on cloud acquisitions between 2011 and 2018. SuccessFactors (cloud HR). Ariba (procurement network). Concur (travel and expense). Hybris (e-commerce). Fieldglass (contingent workforce management). Each deal bought SAP a cloud asset with an existing subscription revenue base, allowing the company to report rapidly growing cloud revenue even as the core ERP installed base remained stubbornly on-premise.
We intend to be the cloud company powered by HANA.
The caution: SAP did not force its on-premise customers to migrate. It offered S/4HANA in cloud, on-premise, and hybrid deployment models, allowing customers to choose their pace. It extended mainstream maintenance for its legacy Business Suite through 2027, then extended it again to 2030 — a series of deadline postponements that reflected the reality that most large SAP customers simply could not migrate quickly. The migration is not an upgrade; it is a reimplementation. ERP consultants estimate that a typical S/4HANA migration for a Global 2000 company takes two to four years and costs $50 million to $500 million, depending on complexity and customization.
McDermott's tenure ended in October 2019, when he abruptly resigned — days after delivering strong quarterly results — to become CEO of ServiceNow. His departure was sudden enough to generate weeks of speculation, but the transition was remarkably clean. SAP appointed a dual-CEO structure: Christian Klein, a 39-year-old SAP lifer who had joined the company as a student intern, and Jennifer Morgan, an American executive who had led SAP's cloud business. The dual structure lasted barely six months; Morgan left in April 2020, and Klein became sole CEO at the age of 40, making him the youngest chief executive of a DAX-30 company.
Klein's Razor
Christian Klein is not the kind of leader who generates magazine profiles or keynote ovations. He is quiet, precise, and operationally ruthless in the manner of a man who has worked at exactly one company his entire adult life and understands its machinery at the molecular level. Born in 1980 in Walldorf — the same small town where SAP was founded eight years before his birth — he grew up in the company's shadow, joined as an intern in 1999, and rose methodically through controlling, corporate finance, and eventually the COO role. He is the institutional product of the thing he now leads.
His defining strategic move came in January 2024, when SAP announced a restructuring program affecting approximately 8,000 roles — roughly 7% of the workforce — primarily in areas being reshaped by artificial intelligence. The restructuring charge was approximately €2 billion. The message was blunt: SAP would redirect resources from legacy on-premise support and sales into cloud and AI development. Investors, who had grown accustomed to SAP's incremental approach, responded with enthusiasm; the stock rose sharply and continued climbing throughout 2024, eventually pushing SAP's market capitalization past €280 billion and making it, briefly, the most valuable company in Europe.
The restructuring was one edge of Klein's razor. The other was the hard deadline: in 2024, SAP began making clear that the 2027/2030 end-of-maintenance dates for legacy ERP were real, and that customers needed to commit to their S/4HANA migration paths. The carrots were RISE with SAP (a bundled cloud migration offering launched in 2021) and GROW with SAP (a cloud-native package for midmarket companies launched in 2023). The stick was the ticking clock.
The results, as of FY2024, were striking. Cloud revenue reached €17.1 billion, up 27% year-over-year, with the cloud backlog — committed but not yet recognized future revenue — growing even faster. The current cloud backlog stood at €18.1 billion, up 29%. S/4HANA cloud revenue specifically grew 72% year-over-year. Total revenue was €31.2 billion. Operating profit, adjusted for the restructuring charge, expanded meaningfully. SAP guided for €36 billion or more in cloud revenue by 2025, with an ambition to reach a cloud gross margin above 80%.
Business AI and the Joule Bet
Every enterprise software company in 2023 and 2024 was compelled to articulate an AI strategy, and most of these articulations were marketing exercises wrapped in press releases. SAP's AI play — branded Joule, a generative AI copilot embedded across the SAP application suite — is more credible than most, for a structural reason: SAP sits on the transactional data of 300,000+ enterprises. It knows what companies buy, from whom, at what price, in what quantities, on what terms, with what payment patterns. It knows their headcount, their compensation structures, their travel expenses, their inventory levels, their production schedules. This is not internet browsing data or social media sentiment — it is the operational substrate of the global economy.
The AI opportunity for SAP is less about building foundation models (SAP partners with Microsoft, Google, Anthropic, and others for underlying LLM capabilities) and more about applying AI to the domain-specific data trapped inside SAP systems. Predictive analytics for supply chain disruptions. Automated invoice matching. Intelligent spend analysis. Dynamic workforce planning. These use cases are less glamorous than chatbots or image generators, but they are precisely the kind of narrow, high-value, enterprise-grade applications where the data advantage is decisive and the willingness to pay is high.
In October 2024, SAP announced that Joule was being embedded into over 100 AI-powered scenarios across its cloud portfolio. More significantly, SAP began introducing AI-based pricing: premium AI capabilities packaged as add-ons to existing cloud subscriptions, in the classic add-on model the St. Gallen Business Model Navigator would recognize instantly — the core offering priced competitively, with numerous AI extras driving the final price up. As Oliver Gassmann, Karolin Frankenberger, and Michaela Csik documented in
The Business Model Navigator, this add-on pattern is one of SAP's foundational business model strategies, and AI represents its latest — and potentially most lucrative — expression.
AI is the next layer of our cloud platform. Customers don't want generic AI. They want AI that understands their business processes, their data, their industry. That is what SAP can deliver.
The Ecosystem as Moat
If the integrated data model is SAP's first moat and process entrenchment is its second, the ecosystem is its third — and arguably the most durable. SAP's partner network encompasses more than 25,000 companies: systems integrators, consulting firms, technology vendors, independent software vendors building on SAP's platform, and resellers. The global SAP consulting market alone is estimated at over $200 billion annually — a figure that dwarfs SAP's own revenue by a factor of six.
This ecosystem creates a self-reinforcing cycle that the St. Gallen framework would describe through multiple overlapping patterns. Accenture employs roughly 75,000 SAP-certified consultants. Deloitte's SAP practice is among its largest globally. These firms have invested billions in SAP-specific training, tooling, and methodology. Their incentive is not merely to sell SAP but to ensure that SAP remains the standard, because their own revenue depends on it. They lobby for SAP within client organizations, recommend SAP in competitive evaluations, and build proprietary accelerators that deepen the integration. This is not a channel strategy. It is a symbiosis — and one that competitors find nearly impossible to replicate.
The ecosystem also includes SAP's Business Technology Platform (BTP), which allows third-party developers to build extensions and custom applications that run on SAP's cloud infrastructure and integrate natively with S/4HANA. The BTP strategy is a platform play: by encouraging partners and customers to build on SAP, the company increases switching costs while simultaneously generating cloud consumption revenue. As of 2024, BTP had become one of SAP's fastest-growing cloud segments, though the company does not break out specific revenue figures.
The European Champion Problem
SAP occupies a singular position in European technology: it is the continent's most valuable technology company, and one of the very few European software firms with genuine global market leadership. This status brings both advantages and burdens. SAP benefits from European data sovereignty concerns — a growing number of enterprises, particularly in the EU, prefer to run their most sensitive business data on a European-headquartered platform rather than on American hyperscalers. Post-Snowden, post-Schrems II, post-GDPR, the compliance advantages of a Walldorf-based company are real.
But the European champion burden is equally real. SAP faces constant pressure from EU regulators on everything from data privacy to competitive practices. Germany's co-determination laws give employee representatives seats on SAP's supervisory board, creating governance dynamics that American shareholders sometimes find opaque. The company's ability to execute the kind of aggressive, move-fast-and-break-things restructuring that American tech firms pursue with impunity is constrained by European labor protections. The 2024 restructuring — 8,000 roles — required extensive works council negotiations.
And then there is the talent war. SAP competes for engineers against Google, Microsoft, Amazon, and a phalanx of well-funded startups, all of which offer compensation packages that SAP historically did not match. Under Klein, SAP has become more aggressive on equity compensation and has expanded its engineering centers in locations like Berlin, Munich, Bangalore, and Palo Alto. But the cultural challenge remains: SAP is a company where "move fast and break things" is anathema to the ethos of engineering reliability that enterprise customers demand.
The Patient Migration
The single most important metric in SAP's universe today is the S/4HANA adoption curve. Of SAP's roughly 437,000 customers, approximately 34,000 had committed to S/4HANA by the end of 2024 — a number that is growing rapidly but still represents a fraction of the installed base. The vast majority of SAP's customers are still running legacy ERP systems (ECC 6.0, Business Suite, or earlier versions) that SAP will eventually stop supporting.
This creates an extraordinary revenue tailwind — a $100 billion+ migration wave that will play out over the next decade. Every legacy customer that moves to S/4HANA Cloud represents a transition from a one-time license and annual maintenance model to a higher-value, multi-year cloud subscription. SAP estimates that the average revenue per customer increases substantially in the cloud versus the on-premise model, driven by higher subscription fees, consumption-based pricing for platform services, and AI add-ons.
But the migration also creates risk. Customers who have delayed for years may decide this is the moment to evaluate alternatives — Workday for HR, Oracle Cloud for financials, ServiceNow for workflows. The migration window is the one period in which SAP's lock-in weakens, because the customer must reimagine their architecture regardless. SAP's competitors know this, and they are investing aggressively to intercept the installed base at the point of maximum vulnerability.
Klein's answer is the RISE with SAP program, which bundles the migration into a single contract covering infrastructure, application licensing, business process intelligence tools, and migration services. RISE is, in essence, an attempt to make the migration decision easy enough — and the alternative evaluation painful enough — that customers default to staying on SAP. Early results suggest it is working: RISE with SAP's annual recurring revenue has grown consistently since its 2021 launch, though SAP does not disclose exact figures.
The Walldorf Paradox
There is a paradox at the heart of SAP that no amount of cloud revenue growth can fully resolve. The company that built the most deeply embedded enterprise software in history — software that runs payroll, closes books, manages supply chains, and processes trillions of dollars in transactions — has spent the past decade in a constant state of self-disruption, cannibalizing its own most profitable business line (on-premise maintenance) to build a cloud future that Wall Street values more highly but that is, in many ways, less defensible. On-premise SAP was a fortress: impossible to rip out, generating decades of annuity-like maintenance revenue, supported by an ecosystem that had no incentive to propose alternatives. Cloud SAP is better software — faster, more flexible, continuously updated — but it is also more competitive. In the cloud, switching costs are structurally lower. The data is more portable. The implementation is less traumatic. The alternatives are more mature.
SAP's bet is that the depth of its functional coverage, the breadth of its industry expertise, and the gravitational pull of its ecosystem will maintain its dominance even as the moat transitions from process entrenchment to platform economics. It is a bet on operational depth over architectural elegance, on the accumulated complexity of global business processes over the clean-sheet simplicity that startups offer. It is a bet that the world's largest companies will always need software that understands the difference between a revenue recognition event under IFRS 15 and ASC 606, across seventy-three jurisdictions, in real time.
On a February afternoon in 2025, SAP's market capitalization hovered around €280 billion. The company had overtaken ASML to become the most valuable firm listed in Europe. In Walldorf, a town still small enough that you can walk from SAP's campus to the train station in twenty minutes, the lights in Building 1 — where Hopp and Plattner first wrote R/1 code — were still on.
