End of Abundance in Tech

1. The Era of Abundance in Tech is Unsustainable

Organizations have been spending increasingly large amounts of money on technology over the past twenty or so years.

Skyrocketing costs. Global technology spending is projected to reach nearly $5 trillion, an 81% increase from 2005 to 2023, largely driven by the shift to cloud-based services. This "era of abundance" has led to a pervasive belief in unlimited resources, but the reality is that cloud migration often results in "sticker shock" due to utility-based, pay-for-what-you-use models. Many technology leaders don't fully grasp the Total Cost of Ownership (TCO) of these platforms, leading to uncontrolled expenses.

Hidden inefficiencies. The primary way organizations currently manage these increased costs is by reducing operational expenses like server count or storage. However, successful businesses amass more customers and data, necessitating more storage and servers unless applications and processes are optimized. The prevailing mindset often overlooks the potential for significant savings—millions, if not tens of millions of dollars—by understanding workload analytics and identifying inefficiencies within technology platforms, particularly in code and processes.

A new paradigm. The author argues that the technology industry needs a fundamental shift in thinking. Instead of merely focusing on keeping systems running, leaders must prioritize technology efficiency and financial transparency. This involves understanding the true costs within applications and services to pinpoint the costliest processes, thereby ending the era of abundance and ushering in an era of efficiency where value is realized from every IT investment.

2. Technology Health Lacks Clear Definition and Measurement

The truth is no one really knows for sure how healthy his or her technology environment is today and how much it will be able to support tomorrow.

Subjective interpretations. Just as different medical specialists might offer varying diagnoses of human health, various tech leaders (CDO, CIO, CTO, CFO) interpret the "health" of a technology environment differently, based on their expertise and the specific data they examine. This lack of a unified, objective measure means that assessments are often inconsistent and fail to provide a holistic view of system well-being or future readiness.

Beyond availability. Current methods for measuring technology health, which haven't significantly evolved in twenty years, primarily focus on availability (uptime). However, true health encompasses more:

• Stability: The system's strength to endure disruption, its reliability, and predictability under stress.
• Performance: Not just functionality, but optimal and efficient operation, minimizing waste.
• Capacity: Whether there are enough resources, avoiding both under- and over-provisioning.
  Without standardized definitions and metrics for these, organizations cannot accurately gauge their systems' true state or the impact of changes.

Actionable insights are missing. While many tools provide data visualizations and dashboards, they often lead to information overload rather than actionable insights. They identify problems but rarely prescribe specific, prioritized actions (e.g., "change X now and Y next to achieve Z"). This gap prevents technical teams from knowing what steps to take to improve systems and how those actions align with key performance indicators (KPIs) or business objectives.

3. Efficiency is the Most Critical, Overlooked KPI

Ultimately it comes down to one thing that influences and impacts just about everything else we’re talking about in this book—efficiency.

The forgotten imperative. In the past, developers were forced to write efficient code due to limited memory. Today, with seemingly endless cloud capacity, the focus has shifted to rapid feature development, often at the expense of efficiency. It's easier to add hardware (CPU, memory) to speed up slow code than to fix the underlying inefficiencies, leading to a "if it's not broken, don't fix it" mentality that perpetuates waste.

The "True MPG" for tech. Just as car buyers consider miles per gallon (MPG) as a standard for fuel efficiency, the technology industry desperately needs a "true MPG" measure for code and application efficiency. Such a standard would allow:

• Assessment of applications during development.
• Identification of inefficient existing applications in production.
• Informed decisions about migrating applications to the cloud, where inefficient code incurs higher costs.
  This would shift the focus from mere functionality to sustainable, cost-effective performance.

Compounding costs of inefficiency. Inefficient code, even if seemingly minor, can have a compounding negative impact over an application's typical 10-20 year lifespan. It drives up hosting, support, and troubleshooting costs, and consumes more resources as data and user volumes grow. This "Total Cost of Code" is rarely considered upfront, leading to tens of thousands, if not millions, of dollars in unforeseen expenses over time. Prioritizing efficient code development and optimization is crucial for long-term financial health and resource conservation.

4. Financial Transparency is Essential but Absent

The majority of organizations are struggling with the increasing costs of technology and the lack of financial transparency into technology services.

The CFO's blind spot. CFOs often lack granular visibility into technology costs. When cloud expenses surge, they typically only see details at the resource or server level, unable to drill down into specific applications or processes to identify the root cause of spending. This disconnect means finance teams struggle to keep pace with the rapid, complex spending decisions made by engineers, leading to budget overruns and a lack of accountability.

Beyond the balance sheet. Financial transparency in technology extends beyond traditional accounting. It requires understanding:

• The true cost of technology (including operational and maintenance expenses).
• The value that technology brings to the business.
• Opportunities for "cost takeouts" through efficiency improvements.
  Without this holistic view, organizations cannot effectively manage non-financial assets like code and data with the same rigor applied to financial information, hindering long-term value creation and eroding trust.

A new framework for accountability. The author advocates for a shift in mindset, moving from simply tracking project costs to understanding the Total Cost of Ownership (TCO) of applications, code, and data throughout their lifecycle. This involves instrumenting financial metrics at the code level, forecasting costs annually, and identifying potential savings from optimization. Gartner's advice to "run technology like a business" by shifting budget planning to include technical, business services, and investment views, along with robust benchmarking, is a step in the right direction for achieving this much-needed transparency.

5. Technology's Environmental Footprint Demands Attention

What’s missing from this conversation is the degree to which technology is contributing to the environmental crisis the world is facing.

The hidden cost of abundance. While ESG (environmental, social, and governance) standards are gaining traction, the technology industry's massive environmental impact is often overlooked. The exponential growth of data—2.5 quintillion bytes created daily, projected to reach 200 zettabytes by 2025—requires enormous energy consumption. US data centers alone consume 73 billion kilowatt-hours annually, equivalent to six million homes, contributing significantly to carbon emissions and e-waste.

Green data centers are not enough. Major tech companies like Microsoft and Google are investing heavily in "green data centers" powered by renewable energy, aiming for carbon-free operations. While commendable, this is only part of the solution. If the applications running within these centers are inefficient and wasteful, consuming excessive CPU cycles and resources, the overall environmental benefit is diminished. The author argues that relying solely on green infrastructure creates a "false sense of comfort."

Towards "green-certified applications." To truly address technology's environmental impact, the focus must extend beyond infrastructure to the applications themselves. A rating system for "green-certified applications" is needed, similar to LEED certifications for buildings. This would involve:

• Measuring the carbon footprint of applications.
• Optimizing code and processes to reduce resource consumption.
• Implementing data lifecycle management to control data growth.
• Extending the lifespan of hardware.
  By making applications more efficient, organizations can significantly reduce power, cooling, and resource demands, translating into substantial carbon savings and contributing to a more sustainable future.

6. Total Cost of Ownership (TCO) Must Include Code and Data

What is the Total Cost of Ownership of the application, code, or process?

Beyond the sticker price. Just as a car's TCO includes fuel, maintenance, and depreciation beyond its purchase price, technology TCO must encompass more than initial acquisition. Many vendors underestimate the resources needed, leading to systems failing to meet expectations within a year, necessitating costly upgrades. This lack of foresight compounds expenses over an application's lifespan, often for 10-20 years.

The compounding cost of code and data. Two critical, often overlooked, components of TCO are:

• Total Cost of Code: Inefficient code, even if minor, consumes more resources (CPU, I/O, memory) with every execution. Over years, this translates to millions in additional hosting, licensing, and support costs. Optimizing the top 5% of inefficient processes can yield 40-80% efficiency gains.
• Total Cost of Data: Beyond storage costs, data growth incurs "soft costs" for management, maintenance, querying, and backup. Data hoarding, where old data is never archived or purged, significantly impacts performance and operational expenses.
  These costs are dynamic and grow exponentially with business and data expansion, yet are rarely factored into initial TCO estimates.

FinOps and the TCO gap. While FinOps aims to bring financial accountability to cloud spending, current tools often stop at the resource level, failing to drill down into the code and application inefficiencies that drive costs. A comprehensive TCO model must integrate:

• Initial Costs: Installation, customization, migration, training, licenses.
• Ongoing Costs: Maintenance, subscriptions, hardware (compute, memory, storage, bandwidth).
• Cost Takeouts: Savings from rightsizing, process optimization, workload rebalancing, downgrading services, and data lifecycle policies.
  This approach provides a realistic financial view, enabling proactive cost management and strategic investment decisions, rather than reactive budget adjustments.

7. Shift from Reactive Problem-Solving to Proactive Optimization

No one is thinking, Could we do what we are doing faster, better, smarter (i.e., more efficiently)?

The "firefighting" trap. The prevailing culture in IT is often reactive, focused on preventing outages and fixing immediate problems. This "firefighting" mentality means that opportunities for proactive optimization—making systems faster, better, and smarter—are consistently overlooked. Unless a system breaks or budget cuts are mandated, teams rarely revisit existing code or processes to improve efficiency, perpetuating a cycle of waste and missed opportunities.

Prioritizing impact over effort. To break this cycle, organizations must adopt a strategic approach to optimization, prioritizing actions based on their potential impact versus the effort and risk involved. This involves:

• Configuration changes: Optimizing OS, platform, or application settings.
• Resource adjustments: Rightsizing servers or service offerings.
• Process refinement: Auditing and disabling legacy processes, changing timing, or reducing maintenance.
• Code and data optimization: Improving indexing, archiving data, optimizing code, and refactoring data structures.
  These actions, when measured before and after implementation, demonstrate tangible value beyond just fixing a problem.

Reframing tech debt. Technical debt, often seen as a negative consequence of "quick and dirty" solutions, can be reframed as "continuous product health." Instead of being an afterthought, addressing tech debt should become an everyday practice, integrated into development sprints. Empowering engineers to prioritize "fixing work" and "investment work" alongside new features, and providing them with the tools and incentives to identify and optimize inefficient code, is crucial for fostering a proactive, efficiency-driven mindset.

8. Translate Technical Wins into Business Value

Today people are communicating wins in technology terms to a business audience that doesn’t understand technology.

The language barrier. Technical teams often celebrate "wins" in terms of reduced I/O, faster runtimes, or lower CPU cycles. While meaningful to engineers, these metrics are often incomprehensible to business leaders and CFOs, who need to understand impact in terms of dollars and cents, customer experience, or strategic advantage. This communication gap prevents technology's true value from being recognized and prioritized at the executive level.

Quantifying impact in business terms. To bridge this gap, technology wins must be translated into business outcomes. For example, instead of reporting a 65% reduction in I/O, communicate:

• "Total report duration reduced from 35 days to 4 days annually, saving 31 days of user waiting time."
• "Estimated annual cost savings of $31,807 in cloud hosting, representing a five-year savings of $159,035."
  This approach provides context, demonstrating the long-term financial and operational value of optimization efforts, making it easier for business stakeholders to approve future investments in efficiency.

Building a feedback loop. A systematic feedback loop is essential to continuously demonstrate value. This involves:

• Baselinining systems: Capturing performance, runtime, and metadata before changes.
• Measuring impact: Quantifying "before and after" results for every optimization.
• Communicating value: Presenting these results in business-centric language, focusing on 1, 3, and 5-year impacts.
• Integrating into roadmaps: Ensuring optimization efforts are consistently incorporated into product development and budget planning.
  This process transforms developers into business stakeholders, aligning their work with organizational goals and fostering a culture where efficiency is directly linked to profitability and customer satisfaction.

9. Institutionalize Efficiency Through Culture and Tools

We need to build a feedback loop into our tools so the savings or efficiencies that are accrued as a result of the changes get communicated back to business stakeholders.

Beyond individual effort. Achieving widespread efficiency requires more than isolated efforts; it demands a cultural shift and the right tools. The "save a gram a month" philosophy from Ford Motor Company illustrates how small, consistent efforts across an organization can lead to massive cumulative savings. In tech, this means every engineer, developer, and administrator must be empowered and incentivized to identify and act on inefficiencies.

Empowering engineers with accountability. A key strategy is to create direct accountability loops, where developers interact directly with clients or see the real-time impact of their code on costs and performance. This fosters a sense of ownership and motivates them to write efficient code. Furthermore, job reviews should include metrics on code efficiency and resource consumption, not just lines of code or feature delivery. This encourages a mindset of "continuous product health" rather than just meeting functional requirements.

The role of advanced analytics and AI. To scale efficiency efforts across hundreds or thousands of applications and servers, organizations need advanced tools. These tools should:

• Provide an "efficiency score": A standardized metric for code and application efficiency.
• Automate impact measurement: Quantify the financial and operational savings of optimizations in real-time.
• Leverage AI/ML: Identify anomalies, forecast future workloads, and recommend optimal actions based on past successes.
  By integrating these capabilities, organizations can move beyond manual, subjective assessments to a data-driven, proactive approach that systematically drives efficiency, reduces costs, and ensures long-term sustainability.