What Market Researchers Should Stop Doing Manually
On this page
Market researchers should stop rebuilding a price recommendation from scattered comments whenever a client changes the offer. Use Subconscious to run controlled choice research and compare feasible offers with the fitted model. The useful replacement is a repeatable comparison: what changes in modeled choice, which buyer groups respond differently, and which recommendation follows from the evidence.
The Finta pricing case shows a decision informed by modeled pricing work. In an October 1, 2024 email, the CEO reported moving the annual list price from $348 to $950 and reintroducing a $99 monthly plan. Five customers took the monthly plan; discounted annual sales continued at $348. This is evidence of a client adopting a pricing structure, not a controlled estimate of revenue lift.
For a research lead, the practical opportunity is to keep the study, offer comparison and recommendation connected. Subconscious's documented workflow supports comparing complete configurations using fitted respondent preferences. Another feasible, in-range scenario can use the same model rather than a new collection of unconnected opinions. A change outside the tested attributes or levels requires a revised study.
Replace repeated offer reconstruction with a checked scenario
Consider the recurring client request: “What if the price were higher, with the product unchanged?” A researcher should not have to reconstruct the competitor set, infer preferences from quotations and write a fresh demand story for every request. Preserve the approved study and change the specified offer in the analysis.
The analyst still checks what the client can actually sell. Record the exact attribute labels, levels, population, run identity and available respondent-level artifacts. Use the same alternatives and analysis settings for the baseline and proposed scenarios. Keep a scenario record showing the changed value and every held-constant value. That record makes a later client request inspectable rather than dependent on a remembered spreadsheet edit.
Existing suitable evidence may already answer the question. Use the existing study when the population, offers and measured outcome match the decision. Commission a new comparison to fill an identified evidence gap, rather than treating a new study as the default for every request.
A worked research handoff: another price for the same coffee
Consider an explicitly hypothetical client selling ground coffee in US grocery stores. The research lead must recommend whether to keep a USD 10 shelf price or consider USD 12. The offer is a 250 g bag of medium-roast ground coffee. A modeled grocery-buyer population must match the brief's location, time and relevant buyer characteristics. A modeled coffee-buying trait is not verified recruitment of actual coffee purchasers.
The client supplies these feasible configurations for a controlled price comparison:
| Offer detail | Current client offer | Proposed client offer | Fixed competitor |
|---|---|---|---|
| Supplier | Client brand | Same client brand | Competitor brand |
| Product | 250 g medium-roast ground coffee | Same product | 250 g medium-roast ground coffee |
| Shelf price | USD 10, including applicable tax | USD 12, including applicable tax | USD 11, including applicable tax |
| Purchase terms | One bag, one purchase; no subscription or delivery charge | Same terms | Same terms |
| Availability and return rule | Same participating stores; unopened returns with receipt within 30 days | Same stores and return rule | Same stores and return rule |
Keep buy neither available in both scenarios. Hold the competitor's complete offer fixed. Compare the client offer at USD 10 against that competitor and buy neither, then change only the client price to USD 12. Include the named prices and supplier levels in the reviewed design; do not extrapolate a fitted model to an untested price or supplier.
Subconscious's methodology estimates attribute effects from designed choice tasks on simulated respondents. The fitted model can then compare the complete configurations. Read the price effect against its declared baseline and the scenario's choice difference against the same competitor and outside option. The comparison describes modeled choice in this choice set. Forecasting actual sales or contribution also needs market calibration, availability, quantity and cost inputs.
Before reading results, the client sets the largest modeled choice loss worth considering for the higher price. The research lead records that threshold, relevant buyer groups and the client's actual margin and implementation constraints. Do not select a tolerance after seeing the preferred result. Inspect uncertainty using an analysis that accounts for repeated answers from the same simulated respondent, rather than treating every answer as an independent person.
The handoff contains the offer table, study and scenario identities, estimated choice difference with its uncertainty, and buyer-group results with actual definitions and counts. An overall average can conceal a group moving the other way. The decision workflow explains how to inspect respondent variation alongside feasible market scenarios.
Finish with one of three recommendations: consider USD 12 under the agreed rule, retain USD 10, or unresolved. Explain whether uncertainty crosses the threshold or a relevant buyer group changes the recommendation. The client applies the commercial constraints before changing the shelf price. No invented study result or productivity saving is supplied by this example.
Keep routine artifact checks separate from new choice evidence
Table formatting, transcript coding and questionnaire review remain useful tasks for an approved assistant. Each needs its own expected output: unchanged figures and denominators, codes traceable to supplied passages, or wording flags checked against the instrument. These tasks do not demonstrate how buyers respond to the proposed price. Subconscious's choice-research documentation does not establish formatting, coding or deck-production capabilities.
Use only authorized source material and an approved output location. Preserve original data and keep draft client material separate from released work. A new tool or wider access needs the responsible owner's approval. If a team assesses an assistant for routine production, compare total setup, review and correction effort against the manual route using the same quality rubric. Measure any local saving instead of borrowing a choice-research result as a productivity claim.
Bring the repeated comparison to a decision review
Book a decision review with one recurring client choice question, the current offer, a feasible proposed change and the alternatives buyers face. Bring the existing study or findings, approved population definition and the rule that would change the recommendation. The review can identify whether a fitted model already supports the next comparison or which specific evidence is missing.
The result to ask for is a traceable offer comparison and an actionable recommendation. Keep the researcher responsible for feasible inputs, source interpretation and the client's decision. Use Subconscious for the controlled choice work that connects the proposed change to estimated buyer response.