Enterprise marketing teams now operate technology stacks that rival the complexity of their company's ERP environments. A single B2B organization might run Oracle Eloqua for demand generation, Salesforce for CRM, a CDP for identity resolution, six intent data vendors, and a handful of point solutions for webinars, chat, and direct mail. Yet according to a recent Q&A published by Demand Gen Report with eClerx's Scott Houchin, 78% of these stacks fail to meet their stated business goals. The instinct is to blame the tools. The real diagnosis points somewhere less comfortable: the operating models that govern how those tools connect to business outcomes.
This statistic should alarm every CMO and VP of marketing operations reading it. Not because it is surprising (most practitioners feel this in their daily work) but because it quantifies an organizational dysfunction that has persisted for over a decade, through multiple waves of platform consolidation, cloud migration, and now AI adoption. The martech industry has added roughly 3,000 tools since 2020, according to Scott Brinker's annual Martech Landscape. Yet the activation gap between insight and action that Houchin describes has only widened.
1. Historical context
The first generation of marketing automation platforms (Eloqua, Marketo, Pardot) emerged in the late 2000s with a straightforward promise: automate repetitive campaign tasks, score leads, and pass qualified contacts to sales. The operating model was simple because the scope was narrow. One platform, one database, one workflow engine.
By 2015, the stack had already started to fragment. The rise of account-based marketing introduced new platforms like Demandbase and Terminus alongside the existing automation layer. CDPs arrived to solve identity resolution problems that the automation platforms were never designed to handle. Intent data vendors layered on signals from the open web. Each addition solved a specific problem in isolation but introduced new integration debt.
Gartner's 2019 Marketing Technology Survey found that marketers reported using only 58% of their martech stack's capabilities. By 2023, that figure had dropped to 33%. The stack was growing; utilization was shrinking. This is the paradox Houchin's 78% failure figure crystallizes. Organizations kept buying solutions to problems that were, at their root, operational and strategic rather than technological.
The pattern repeated with each hype cycle. Social listening tools were bolted on without clear workflow connections to campaign execution. Conversational marketing platforms generated leads that sat in queues disconnected from the scoring models in Eloqua or Marketo. Data enrichment services pumped new fields into CRMs that no one had mapped to segmentation logic. Every tool worked. The system did not.
This history matters because it explains why the current moment, with AI agents and copilots arriving across every platform, carries the same risk. Without an operating model that defines how tools connect to strategy, how data flows between systems, and who owns each handoff, the addition of AI will simply add a new layer of expensive underutilization on top of the existing stack.
Source: Gartner Marketing Technology Survey 2020, 2022, 2023
"Marketing technology is growing faster than organizations' ability to use it."
2. Technical analysis
Houchin's diagnosis centers on what he calls the "activation gap": the distance between having a customer insight and being able to act on it in a coordinated way across channels. This gap has specific technical anatomy worth dissecting.
The data foundation problem
Most enterprise stacks suffer from fragmented identity graphs. A contact might exist in Salesforce with one email, in Eloqua with a slightly different spelling of their company name, and in the CDP with a cookie-based anonymous profile that has not yet been stitched. When these records do not resolve to a single identity, every downstream action (scoring, routing, personalization, suppression) operates on partial information.
The problem compounds at the account level. ABM strategies require account-level signals (intent scores, engagement aggregates, technographic data) to be resolved against contact-level actions. When the data quality layer is unreliable, account scoring becomes a fiction. Sales teams learn to ignore marketing-qualified accounts because the signal has been wrong often enough to destroy trust. As we explored in our analysis of how bad data undermines CRM-email convergence, this is not a new problem. It is a persistent one that compounds with every new data source added to the stack.
The workflow disconnection problem
Even when data is clean, most stacks lack connected workflows that span systems. A typical example: an intent data vendor flags that a target account is surging on a relevant topic. That signal lands in a dashboard. A demand gen manager sees it 48 hours later, creates a manual list, builds a campaign in the automation platform, and launches an email. By the time the contact receives the message, the buying window may have shifted.
The technical gap here is not intelligence. It is orchestration. Connected workflows require event-driven architectures where a signal in one system triggers an automated action in another, with appropriate governance and human-in-the-loop checkpoints. Most stacks are built on batch integrations (nightly syncs, CSV uploads, scheduled API pulls) that are structurally incapable of real-time activation.
The measurement disconnection
The third layer of the activation gap involves measurement. When campaign execution spans multiple tools (ads in one platform, email in another, webinars in a third, sales outreach in a fourth), attribution becomes an exercise in reconciliation rather than analysis. Each system claims credit for its own touchpoints using its own model. No single view of the buyer journey exists. As a result, optimization decisions are made on incomplete evidence, and budget allocation follows inertia rather than performance signals. We examined the structural roots of this problem in our piece on the analytics architecture gap.
3. Strategic implications
The 78% failure rate has consequences that extend well beyond the marketing department. Three stand out for enterprise leadership.
Revenue leakage through handoff failures
The most immediate cost of a disconnected stack is revenue leakage at the marketing-to-sales handoff. When lead scoring models rely on incomplete behavioral data (because tracking is fragmented across systems), they produce false positives and false negatives at roughly equal rates. Sales receives leads that are not ready to buy and never sees leads that are. A 2024 Forrester study on B2B revenue waterfall efficiency found that organizations with unified data foundations converted marketing-qualified leads to sales-accepted leads at 2.4x the rate of those with fragmented stacks.
For a company generating 10,000 MQLs per quarter, even a modest improvement in conversion rates at this handoff translates to millions in pipeline. The fix is not a better scoring algorithm. It is a connected lead scoring infrastructure that ingests behavioral, firmographic, and intent data from a single resolved identity.
Strategic planning paralysis
When measurement is unreliable, planning becomes political. Without trusted performance data, budget allocation decisions default to the loudest voice in the room or the program that has always received funding. This is how organizations end up spending 40% of their demand generation budget on trade shows that produce business cards but no pipeline, while underfunding digital programs that actually move revenue.
A connected stack produces the performance visibility that enables evidence-based strategic planning. Without it, every planning cycle is a negotiation rather than an analysis.
Compounding technical debt
Every disconnected tool added to the stack creates integration debt that makes the next tool harder to connect. Over time, the stack develops what engineers call "accidental complexity": complexity that exists because of how the system evolved rather than because the business problem requires it. This debt slows campaign execution, increases error rates, and makes the operations team a bottleneck rather than an accelerator. Organizations find themselves unable to adopt new capabilities (AI-powered personalization, real-time orchestration, predictive analytics) because the foundation will not support them.
"Companies that try to do everything with their martech stack end up doing nothing well. The key to success is ruthless prioritization of use cases."
4. Practical application
Closing the activation gap requires a structured approach that addresses data, workflows, and governance in sequence. Skipping steps or attempting to fix everything simultaneously is how organizations end up buying another tool to compensate for the tools they already have.
Step one: audit the current state honestly
Most organizations lack an accurate inventory of their martech stack. A 2024 Productiv study found that the average enterprise has 40% more SaaS tools than leadership believes. Begin with a complete enumeration: every tool, every integration, every data flow, every manual process that bridges systems. Map each tool to a specific stage in the revenue process (awareness, engagement, conversion, expansion) and identify gaps and overlaps.
A platform maturity assessment provides the structured framework for this audit. The output should be a system-of-record map showing which platform owns each data object, which direction data flows, and where manual handoffs introduce latency or error.
Step two: establish a trusted data foundation
Before any workflow optimization, the data layer must be reliable. This means resolving identities across systems so that every platform operates from the same understanding of who a contact is and which account they belong to. It means establishing data normalization standards so that "IBM," "International Business Machines," and "IBM Corp" resolve to a single entity. And it means implementing ongoing data quality processes rather than treating data cleansing as a one-time project.
The specific technical approach depends on the stack architecture. Organizations running a CDP may centralize identity resolution there and syndicate resolved profiles to downstream systems. Those without a CDP may use the CRM as the master record with bidirectional sync to the automation platform. The architectural choice matters less than the discipline of having a single, governed source of truth.
Step three: connect execution workflows
With clean data flowing, the focus shifts to connecting execution. The goal is to reduce the latency between a signal (a lead score threshold crossed, an intent surge detected, a form submitted) and the corresponding action (an email sent, a sales alert triggered, an ad audience updated).
Prioritize the highest-value workflows first. For most B2B organizations, these are:
- Lead routing from marketing automation to CRM with complete context (behavioral history, content consumed, intent signals)
- Multi-touch campaigns that span email, ads, and sales outreach with shared audience logic
- Event-triggered nurture sequences that respond to buying signals within hours rather than days
Each connected workflow should be documented, monitored, and governed. Assign ownership for each handoff point. Define SLAs for response times. Instrument the workflow so that failures are detected and resolved before they affect revenue.
Step four: build feedback loops into planning
The final step closes the loop between execution and strategy. Connected measurement requires standardized definitions (what counts as an MQL, what constitutes an opportunity, how attribution credit is distributed) and a single reporting environment that aggregates data from across the stack.
This is where most organizations stall. Building a unified measurement layer requires agreement across marketing, sales, and finance on definitions that have historically been contested. The work is political as much as technical. But without it, the stack remains a collection of tools rather than a system, and the 78% failure rate persists.
5. Future scenarios
Two trajectories are plausible over the next 18 to 24 months, and they diverge sharply.
Scenario one: AI amplifies the activation gap
In this scenario, organizations layer AI agents and copilots onto their existing disconnected stacks. Each platform vendor ships its own AI assistant (Salesforce Einstein, HubSpot Breeze, Oracle Eloqua AI, Adobe Sensei) and each assistant optimizes within its own silo. The email AI optimizes subject lines based on email engagement data alone. The CRM AI predicts deal outcomes based on CRM data alone. The intent data AI surfaces accounts based on web signals alone.
The result is locally optimized, globally incoherent execution. An account might receive a personalized email nurture from marketing while sales AI has already flagged it for direct outreach on a different message. Intent signals might trigger automated ad campaigns for accounts that have already converted. The activation gap does not shrink; it multiplies across AI-driven channels operating without shared context.
This scenario is the default for any organization that adds AI to a stack without first addressing the data and workflow problems described above. It is also the most likely outcome for the majority of enterprises, given current adoption patterns. As we discussed in our analysis of ABM's integration deficit, personalization at scale remains a plumbing problem, and AI does not fix bad plumbing.
Scenario two: AI forces operational consolidation
In the more optimistic trajectory, the demands of AI adoption force organizations to do the hard work of stack consolidation and data unification that they have deferred for years. The logic is straightforward: AI agents require clean, connected data to function. Organizations that want to deploy agentic workflows spanning marketing, sales, and customer success will discover that those workflows cannot operate on fragmented data and batch integrations.
This realization could trigger a wave of stack simplification. Rather than adding tools, organizations may remove them, consolidating onto fewer platforms with deeper platform integrations and investing in the data infrastructure that makes AI activation possible. We may see a resurgence of interest in marketing automation strategy as organizations redesign their stacks around a smaller number of well-connected systems rather than a large number of loosely coupled ones.
The organizations that follow this second path will separate themselves from the 78%. They will move faster, measure more accurately, and convert more efficiently. The gap between leaders and laggards in revenue operations will widen, and it will widen on the basis of operational maturity rather than technology budget.
The platform vendor response
Platform vendors are beginning to recognize the consolidation opportunity. HubSpot has expanded aggressively into enterprise with a unified CRM, marketing, sales, and service platform. Salesforce is pushing Data Cloud as the connective tissue across its ecosystem. Adobe has integrated its Experience Platform across Marketo and the broader Experience Cloud. Oracle continues to position Eloqua within a broader CX suite.
The question for enterprise buyers is whether platform consolidation alone solves the operating model problem. The answer, based on a decade of evidence, is no. Organizations that migrated from best-of-breed stacks to single-vendor suites in the 2018 to 2022 period often found that the integration problems simply moved inside the vendor's ecosystem. Consolidation reduces the number of integration points but does not eliminate the need for coherent strategy, clean data, and governed workflows.
6. Takeaways
-
The 78% failure rate of martech stacks reflects operating model failures, not technology failures. Adding tools without connecting them to strategy, data, and workflows produces expensive underutilization.
-
The "activation gap" between insight and action has three layers: fragmented data foundations, disconnected execution workflows, and unreliable measurement. Each must be addressed in sequence.
-
Revenue leakage at the marketing-to-sales handoff is the most immediate cost of a disconnected stack. Organizations with unified data foundations convert MQLs at significantly higher rates.
-
AI adoption will either amplify the activation gap (if layered onto disconnected stacks) or force the consolidation work that should have been done years ago. The outcome depends on whether organizations invest in data and workflow infrastructure before deploying AI agents.
-
Practical remediation starts with an honest stack audit, moves to data foundation work, then connects execution workflows, and finally builds unified measurement. Skipping steps guarantees failure.
-
Platform consolidation reduces integration points but does not substitute for coherent operating models. Governance, ownership, and cross-functional agreement on definitions matter as much as the technology itself.
-
The gap between revenue operations leaders and laggards will widen over the next two years, and it will widen on the basis of operational maturity rather than technology spending.
