Question 1

What's the cleanest way to test banking kiosk versus teller journeys?

Accepted Answer

The cleanest way to compare a bank's self-service kiosk experience with a human teller experience is to run a controlled test where customers are randomly assigned to each option for the same task. For instance, you might have a group of customers perform a common transaction — like depositing a check or opening an account — one group using the kiosk and a similar group working with a teller. By measuring outcomes such as transaction time, error rates, and customer satisfaction immediately afterward, you isolate the differences attributable to the channel. Random assignment is key here: it ensures that any differences in results (say, kiosk users finished two minutes faster on average, or teller users felt more reassured) are due to the method of service, not because one group had inherently easier cases or more tech-savvy individuals. This A/B test style approach provides clear, unbiased evidence of how kiosks stack up against tellers for key banking journeys.

Question 2

Can AI-driven web scraping replace human analysis in CI?

Accepted Answer

AI-driven web scraping can automate data collection for competitive intelligence (CI), but it can't fully replace human analysis due to context and judgment. An AI scraper will tirelessly gather competitor prices, product changes, or reviews from across the web at scale and speed that humans can't match, giving your CI team a firehose of raw information. However, making sense of that information - filtering signal from noise, interpreting why a competitor made a move, or spotting subtle market shifts - still requires human insight. AI might flag an anomaly (like a sudden price drop), but a human analyst determines if it's a strategic play, a stock clearance, or an error. There's also nuance and sarcasm in online chatter that AI might miss or misread. The sweet spot is using AI to do the heavy lifting of monitoring and initial summarizing, then having analysts apply expertise to validate findings and draw strategic conclusions. This hybrid approach mitigates the risk of blind spots or false leads that a solely AI system might generate.

Question 3

What practices improve adoption of change initiatives in research?

Accepted Answer

Successful adoption of change initiatives in a research organization relies on strong communication, early involvement, and demonstrating value quickly. Start by clearly articulating why the change is happening — whether it's a new methodology, tool, or process — so the team understands the benefits and the urgency. Involve key researchers or stakeholders early (even in planning or pilot phases) to build buy-in and get feedback, effectively making them champions of the change. Provide plenty of training and support materials, and highlight quick wins: for example, showcase a project where the new approach led to faster insights or better outcomes to prove its value. Leadership should visibly back the initiative, modeling the desired behaviors and publicly recognizing team members who embrace the change. Also create forums for staff to ask questions and share experiences as they adapt. By combining transparency, support, and proof of success, you greatly improve the odds that a change initiative will be adopted rather than quietly resisted.

Question 4

When should AMI data calibrate survey-based load insights?

Accepted Answer

AMI data should be used to calibrate survey-based load insights whenever real usage patterns are needed to validate or adjust what people report. Advanced Metering Infrastructure (AMI) provides granular, objective measurements of energy load, which can correct for biases in survey responses (such as recall error or optimism about usage). For example, if customers claim certain peak usage times or conservation behaviors, comparing those survey insights with actual smart meter data helps align the findings with reality_x000D_. In practice, this calibration is most valuable in forecasting or program evaluation: by integrating AMI readings, you ensure that decisions (like grid planning or demand response strategies) are based on accurate, observed consumption rather than solely on self-reported data.

Question 5

Does adding more variables always make a model better?

Accepted Answer

No - after a certain point, adding more variables can actually weaken a model by adding noise or multicollinearity rather than boosting accuracy. While it's tempting to include every possible predictor, irrelevant or redundant variables may make the model fit the quirks of your training data (overfitting) and then perform worse on new data. More variables also mean more complexity, which can lead to unstable coefficients if predictors overlap in what they explain. In practice, the most important factors usually provide the biggest lift in predictive power, and each additional variable has diminishing returns. Data scientists often use techniques like cross-validation or feature importance analysis to decide if a new variable genuinely improves the model. The bottom line: a simpler model with the right variables is typically more robust and interpretable than a bloated model with every data point thrown in.

ReDem

Why choose Us

What is ReDem?

Research Software and Platform Details

Services and Expertise

Research Resources

Market Research FAQs with Expert Insights

What's the cleanest way to test banking kiosk versus teller journeys?

Can AI-driven web scraping replace human analysis in CI?

What practices improve adoption of change initiatives in research?

When should AMI data calibrate survey-based load insights?

Does adding more variables always make a model better?