I replied to every cold email in my inbox for seven days. Instead of behaving like a normal prospect, I asked vendors to explain their methodology, support their claims, provide proof, and respond to requests that did not fit neatly inside a sales sequence. One vendor was asked to write a raccoon accounting haiku before I would take a meeting.
The experiment started as a way to evaluate salespeople. It ended as a study of what outbound systems are optimized to do when a prospect stops behaving predictably.
Research question
What happens to modern outbound sales systems when a prospect responds, asks direct questions, and refuses to follow the expected path toward a booked meeting?
Methodology
For seven days:
- Every cold email received was answered.
- Vendors were asked for methodology, data, proof, or differentiation.
- Some vendors received unusual but harmless requests.
- The prospect stayed responsive.
- Lack of interest was communicated when relevant.
- Vendor and salesperson identities were anonymized.
The analysis below uses only the documentary transcript. It does not claim that the observed conversations represent the entire sales industry.
Research summary
Four recurring behavior patterns emerged:
- Script readers sounded personalized until the prospect asked a question that required a new answer.
- Compliance officers completed every requested task but failed to establish meaningful differentiation.
- Follow-up machines continued escalating sequences without resolving the questions already asked.
- Human adapters temporarily left the script, engaged naturally, and became more memorable, even when their underlying offers remained generic.
Across the experiment, conversations repeatedly drifted toward the same destination: activity, persistence, and booked meetings.
Key findings
Personalization often ended after the first email
A podcast growth company claimed it had audited the podcast, identified visibility problems, and calculated missed viewers. When asked how the calculation worked, the response was a calendar link.
The initial message appeared specific. The follow-up behavior did not support the claim.
Completing every request did not create differentiation
An accounting firm answered the questions, followed the requested process, and wrote the raccoon accounting haiku:
Tiny paw sorting.
Revenue earned not yet paid.
Moon books balanced.
The unusual request made the company memorable. Its methodology did not.
Persistence frequently replaced listening
The most common pattern in the transcript was repeated follow-up activity that did not answer the prospect's questions. One sequence became increasingly passive-aggressive while leaving the original request unresolved.
Human adaptation briefly stood out
A podcast editor played along, answered questions, offered specifics, wrote the haiku, and respected the premise of the experiment. That behavior made him stand out because it appeared responsive rather than scripted.
Even that conversation eventually returned to the same destination: getting a call booked.
The system revealed its optimization target
When the prospect stopped behaving predictably, participants revealed what their processes appeared designed to maximize:
- Activity.
- Persistence.
- Meetings booked.
The transcript found much less evidence of systems optimized for curiosity, comprehension, or deciding whether a meeting made sense.
Behavior taxonomy
Script readers
Personalized opening. Predetermined response after the prospect asks a direct or unexpected question.
Compliance officers
Complete the requested tasks but treat each request as another box to check. High responsiveness, weak differentiation.
Follow-up machines
Continue the sequence regardless of whether earlier questions were resolved. Interpret replies as progress and persistence as the default response.
Human adapters
Temporarily depart from the sequence, engage with the actual request, and respond in a way that feels specific to the conversation.
Why this matters as outbound becomes more automated
AI can write cold emails, follow-ups, personalization tokens, and multichannel outbound campaigns. As polished outreach becomes cheaper and easier to produce, polish itself becomes less distinctive.
The behavior that stands out may be the behavior automation handles worst:
- Understanding what was actually asked.
- Answering before redirecting.
- Recognizing when a prospect is not qualified or interested.
- Departing from the workflow when the conversation requires it.
- Acting like a person rather than treating every response as pipeline progress.
Limitations
This is a qualitative case study based on one person's seven-day experiment. The findings should be read as documented observations, not prevalence estimates.
Dataset and source files
Source
- Documentary: Why Your Best Systems Collapse the Moment They're Tested (YouTube)
- Original article: jessicamalnik.com
The documentary transcript is the sole factual source for this research package. Every claim and category was extracted from the transcript. No outside facts, company identities, or market statistics were added.
