Marketing Performance and Growth Efficiency Analysis
A strategy-oriented analysis of trivago’s marketing efficiency, examining how increased brand investment, traffic mix shifts, and marketplace changes affect sustainable…
Project Overview
This project asks a single, falsifiable question: can AI replace human reps in high-volume, low-ticket B2B phone sales today? It answers the question by synthesizing peer-reviewed empirical evidence rather than vendor claims or opinion. The review is structured to be proven wrong: the proposition is stated as a hypothesis to reject, the criteria that would change the conclusion are fixed before the evidence is examined, and sources are screened for commercial interests that could bias their findings.
Because the author also builds a product in this market, the review is deliberately designed to counter that potential bias. The result is a decision-oriented assessment of what the research currently supports — where AI already matches human performance, where customer acceptance still blocks replacement, and where AI creates measurable value by augmenting reps instead.
Key Features & Components
Falsifiable question, criteria fixed in advance
Frames the proposition the author has a commercial interest in as the hypothesis to reject, with the criteria that would overturn the conclusion set before any evidence was gathered.
Conflict-of-interest source auditing
Screens every candidate source for segment relevance and quality, and excludes any source with a commercial stake in the outcome — on either side of the question.
Capability separated from acceptance
Distinguishes what AI can already do on a sales call from whether customers will accept it — the distinction that decides whether replacement is viable today, not just possible.
Evidence for augmentation over replacement
Synthesizes deployment studies on AI coaching, sequential human–AI funnels, and AI-assisted lead handling to locate where AI creates measurable value.
Explicit scope and limits
States what the evidence base cannot support — small sample, partly dated studies, segment mismatch — rather than overclaiming from it.
NoteFull Report
Read the complete analysis in the Will AI Replace Human Sales Reps? An Evidence Review.
Research Process
Defined the research question as a falsifiable hypothesis and fixed the kill criteria before gathering evidence.
Built a candidate-source collection process, then audited each source for segment relevance, quality (peer-reviewed preferred), and conflicts of interest; excluded any source with a commercial stake.
Reported every figure exactly as its primary source states it, and omitted figures that could not be located in the source.
Authored the report in Quarto, integrating prose and citations into a single reproducible document.
Outcomes & Impact
Turns an open, anxiety-driven question into a defensible, present-tense answer a team can act on this quarter.
Demonstrates bias-aware evidence synthesis despite a limited and partly dated evidence base.
Separates technical capability from customer acceptance, changing how AI investments can be evaluated today.
Provides a decision-oriented basis for sales-technology strategy without overstating what the research proves.
Skills Demonstrated
Evidence synthesis • Research design • Source evaluation & bias auditing • Critical appraisal of empirical studies • Analytical writing • Translating research into strategy
Apply This to Your Business
If you need to evaluate a business question where the evidence is incomplete, conflicting, or easy to overinterpret, I can help structure the analysis, test the assumptions, and turn the findings into a decision.
