Productivity Paradox (Efficiency Paradox)

Table of Contents

  1. Definition
  2. Why it matters for marketing
  3. How it works: why technology investments don't show up in productivity
    1. The marketing ops version: the efficiency paradox
  4. How to utilize the concept
  5. Comparison with related concepts
  6. Best practices
  7. Future trends
  8. FAQs
  9. Related Terms
  10. Listen: related episodes of The Agile Brand
  11. Sources

Also called: Solow paradox, Solow productivity paradox, IT productivity paradox

Definition

The productivity paradox is the observation that large investments in information technology often fail to show up as measurable productivity gains, at least for a long time. Economist Robert Solow summed it up in 1987: “You can see the computer age everywhere but in the productivity statistics.”

In marketing, the paradox shows up when a team keeps adding platforms and AI tools but doesn’t get faster, cheaper, or more effective at producing business results. In Marketing Operations 3.0 (De Gruyter Brill), Greg Kihlström applies the idea to marketing operations under the name efficiency paradox. Tools meant to increase speed often add complexity, and teams end up spending more time managing integrations than improving growth outcomes.

See also: Fauxlution · Total Cost of Ownership (TCO) · Technical Debt · Return on Investment (ROI)

Back to Top

Why it matters for marketing

Marketing has been one of the heaviest buyers of software over the past fifteen years. Scott Brinker’s 2025 Marketing Technology Landscape counted 15,384 martech solutions. Gartner’s 2023 Marketing Technology Survey found that marketers were using only 33% of their martech stack’s capabilities. Spending on tools has clearly outpaced the ability to get value from them.

Generative AI is now repeating the pattern. In August 2025, an MIT NANDA report (The GenAI Divide) found that about 95% of the enterprise generative AI pilots it studied produced no measurable impact on profit and loss. Whatever the exact number, the gap between AI adoption and AI results is the productivity paradox in its newest form.

For marketing ops leaders, the paradox is practical. A CFO looking at rising software costs and flat output will treat the function as overhead to be trimmed. The book argues that cutting the budget doesn’t fix the paradox. Changing what the money is spent on does.

Back to Top

How it works: why technology investments don’t show up in productivity

Economists have offered several explanations for the paradox. Erik Brynjolfsson grouped the early ones in a 1993 paper, and later work added more:

ExplanationWhat it meansMarketing example
MismeasurementGains are real but hard to capture in standard metricsBetter targeting improves customer quality, but reports only track lead volume
LagsBenefits arrive years after the investment, once people and processes adaptA CDP takes 18 months before segments are trusted and used
RedistributionGains for one firm or team come at another’s expense, with no net improvementTwo teams’ automation competes for the same customers’ attention
MismanagementThe technology is used to automate old processes instead of redesigning themAI writes more emails for the same rigid nurture sequence
Complementary investment (the “J-curve”)New general-purpose technologies need large investments in skills, data, and process redesign before gains appearAgents can’t act well until product, pricing, and customer data are machine-readable

The last explanation comes from Brynjolfsson, Daniel Rock, and Chad Syverson’s work on AI and the “modern productivity paradox.” Their argument is that AI behaves like other general-purpose technologies. Productivity often dips or stalls while companies make the less visible investments needed to use it, then rises once those are in place.

Back to Top

The marketing ops version: the efficiency paradox

The book identifies two marketing-specific drivers:

  1. The cost-center trap. When marketing operations is funded as overhead and judged on minimizing spend, its leaders optimize for local efficiency instead of business results. Investment goes to tools that promise labor savings, not to the coordination work that would actually raise output.
  2. The Frankenstack. Rapid tool adoption leaves teams with 30-plus loosely connected systems. People end up “working for the technology,” closing gaps between platforms instead of improving outcomes.

The book’s answer is to move investment “from tools to orchestration”: shared customer data, coordinated decisions across teams, and continuity across the customer lifecycle. As the book puts it, “Spending less on the function is not the same as running the business better.”

Back to Top

How to utilize the concept

  • Audit utilization before buying. Measure how much of each existing platform’s capability is actually in use.
  • Count integration labor. Track hours spent maintaining syncs, reconciling data, and fixing handoffs. Those hours are the hidden cost of the stack.
  • Fund complements, not just tools. Budget for data cleanup, process redesign, training, and governance alongside any new platform or AI capability.
  • Measure outcomes at the system level. Look at cycle time, conversion, retention, and cost per outcome across the whole lifecycle, not just individual tool metrics.
  • Use it in the business case. Frame investments in data and orchestration as the way to unlock value from tools already paid for.
  • Expect a J-curve. Set expectations with leadership that AI programs may show costs before returns, and define the milestones that mark progress.
Back to Top
ConceptCore ideaRelationship
Productivity paradox (Solow)IT investment doesn’t appear in productivity statisticsThe general economic observation
Efficiency paradox (MarketingOps 3.0)Marketing tools meant to speed work up add complexity that cancels the gainsThe marketing operations application
Productivity J-curveProductivity dips before it rises as firms build complementary capabilitiesOne explanation for the paradox, especially for AI
Jevons paradoxEfficiency gains increase total consumption of a resourceDifferent. It explains rising use, not missing gains, though both are about efficiency gains that don’t play out as expected
FauxlutionA change that looks like transformation but doesn’t change outcomesA common cause. Fauxlutions add investment without adding productivity
Technical DebtAccumulated cost of shortcuts in systemsA contributor. Debt in the stack absorbs effort that new tools were supposed to free up
Back to Top

Best practices

  • Treat tool count as a cost, not an asset.
  • Pair every new platform or AI capability with a named process change. Automating an unchanged process is the most common way to recreate the paradox.
  • Retire tools as deliberately as you buy them.
  • Report on reclaimed time and where it went. Hours saved that disappear into more coordination aren’t a gain.
  • Define what “productive” means for the function before measuring it, such as outcomes per dollar or cycle time from signal to action.
  • Don’t judge AI programs only in their first quarter. Do define the leading indicators that show whether the complementary work is getting done.
Back to Top
  • AI as the next test case. Economists and analysts are watching whether generative and agentic AI follow the J-curve pattern or break it.
  • Stack consolidation. Pressure to cut redundant martech is likely to grow as AI agents need consistent data across fewer systems.
  • Outcome-based metrics. Measures such as Return on Agent (ROA) aim to evaluate AI-driven work by the value it creates against its full cost, including governance.
  • Pilot-to-production discipline. Reports of low AI pilot success rates are pushing companies to set clearer criteria for scaling or stopping AI projects.
  • Better measurement. Expect more attention to capturing intangible gains, such as faster decisions and better customer quality, that standard reports miss.
Back to Top

FAQs

What is the productivity paradox? The observation that heavy investment in information technology often doesn’t produce measurable productivity gains, or produces them only after a long delay.

Who coined it? It’s usually traced to economist Robert Solow’s 1987 remark that “you can see the computer age everywhere but in the productivity statistics.” Erik Brynjolfsson’s 1993 paper helped popularize “productivity paradox” as the term.

What is the efficiency paradox in marketing operations? Greg Kihlström’s term in Marketing Operations 3.0 for the way marketing tools meant to increase speed often add so much complexity that teams spend more time managing integrations than improving results.

Was the original productivity paradox ever resolved? Partly. U.S. productivity growth picked up in the late 1990s, and research by Brynjolfsson and Lorin Hitt found that IT investment was linked to productivity gains at the firm level, especially when paired with organizational changes.

Does the productivity paradox apply to AI? Many economists think so. Brynjolfsson, Rock, and Syverson argue that AI, like earlier general-purpose technologies, needs large complementary investments before its gains show up.

What is the productivity J-curve? The pattern in which productivity flattens or dips while companies invest in skills, data, and processes for a new technology, then rises once those investments mature.

How can marketing teams avoid the paradox? Invest in data quality, process redesign, coordination across teams, and governance alongside tools, and measure results across the whole customer lifecycle rather than tool by tool.

What is a Frankenstack? An informal term for a marketing technology stack made of many loosely connected tools, where maintaining the tools takes more effort than they save.

How is this different from the Jevons paradox? The Jevons paradox says efficiency gains can increase total consumption of a resource. The productivity paradox is about gains that don’t appear in the first place.

Back to Top
  1. Fauxlution
  2. Technical Debt
  3. Operational Debt
  4. Total Cost of Ownership (TCO)
  5. Return on Investment (ROI)
  6. Build vs. Buy
  7. Time to Value (TTV)
  8. Diffusion of Innovations
  9. MarketingOps Maturity Model
  10. Return on Agent (ROA)
Back to Top

Read: Micro-productivity can’t be all there is to gain from AI (Martech Futurist)

Back to Top

Sources

Back to Top

Tags:

Was this helpful?

Previous

Porter's Generic Strategies

Next

Project Management Components (PMBOK)