Materials · screening/scout-fractional.md

# Scout report: Fractional and contract track

Sources read: northline-behavioral-scientist.md (posting), northline.md (employer packet), record.md (candidate). Nothing else.

## Northline Advisory: Behavioral Scientist (contract, fractional)

Remote, about 10 hrs/week, six months, renewable, paid hourly.

### 1. What Northline is actually selecting for
The packet is clearer than the posting. The founder says clients "run promotions as experiments in name only" and wants someone who can "make a design defensible before money is spent." So the job is quality control on experimental design: catch bad designs before launch, and check analyses before they reach executives. It is a small firm (five people) selling hourly advice to regional retailers. Being able to speak with authority to client executives matters as much as research depth.

### 2. Jordan's evidence (from record.md) and what's missing
Supports:
- Parcel, 2018 to 2022 (Senior Research Scientist): designed and analyzed product experiments on delivery-time estimates and tipping prompts, and built the company's first holdout process for long-running experiments. This is the most direct evidence of designing and analyzing experiments in the field. Tipping prompts are close to pricing but are not pricing.
- Kestrel Media, 2022 to 2024: owned the measurement plan for product-team experiments. The record says Jordan **did not run** those experiments. This matches the "review analyses" part of the job well.
- Pricing-experiment rehearsal (2025): **simulated**. Language-model agents acted as shoppers to test whether a tiered-pricing design, its sample size and its analysis plan could detect a planted effect. This lines up directly with "rehearse designs before launch." It tells you nothing about real shoppers, and the client's own experiment had not reported results.
- Working paper "Rehearsing experiments with simulated participants" (2025): **working paper**, self-posted, not submitted.
- Methods skills: experimental design, Python/R/SQL, multilevel models.
- Relay caught a configuration error that would have contaminated every condition. This is real evidence of design-QA instincts, but in a language-model testbed, not a retail setting.

Missing or absent:
- No retail, pricing or promotion field experiment with real customers.
- No direct work with client executives. The record shows no consulting or client-facing role apart from the single simulated rehearsal done "before a client ran it." Who Jordan dealt with on that client is not recorded.
- No record of a contracting or hourly rate history.
- Note: the record lists the 2019 Academy of Management talk under "Invited," but it was a competitively selected paper session, not an invitation. Do not describe it as invited. It is not relevant here anyway.

### 3. Materials and submission
A short note and a résumé, emailed to the engagement lead. Rate expectations are optional ("welcome"). Jordan sets the rate. The scout does not.

### 4. Strongest case against applying
Jordan has never run a pricing or promotion experiment with real shoppers and has no recorded experience advising client executives. Both are central to a firm that sells credibility to retail executives. The closest pricing evidence is a simulation and an unsubmitted working paper. If these are presented as anything more, they overstate what Jordan has done. The engagement is also small (about 10 hrs/week, hourly, at a five-person firm), so the payoff may be limited compared with what it costs Jordan in focus.

### 5. Title vs actual work
Minor divergence. The title says "Behavioral Scientist," but the work described in the posting and packet is experimental-design and analysis QA for pricing and promotions. It is methods consulting, not behavioral research or theory. This suits Jordan's methods record better than the title suggests, but Jordan should not expect to do research.

**Verdict: pursue.** Parcel experiment design, Kestrel measurement-plan ownership and the simulated pricing rehearsal fit the "defensible design before money is spent" need. Jordan should frame the rehearsal strictly as simulation and decide on a rate.