Weighted Experience Index (IEP): A Composite Metric for Robust Store Performance Evaluation

Evaluating the performance of stores or sales points with a low volume of survey responses is a challenge. Traditional customer experience metrics like the Net Promoter Score (NPS) —which measures customer loyalty or likelihood to recommend— and the Customer Satisfaction Score (CSAT) —which measures immediate satisfaction after an interaction— can become skewed when the number of responses is too small. A single very satisfied (or dissatisfied) customer can inflate or crash the NPS/CSAT of a store with few responses, yielding an unreliable picture. Indeed, it's been pointed out that NPS is "rarely relevant with a small sample size"; with very few respondents, the figure lacks statistical significance. As a result, performance scores of outlets with low survey participation tend to show high variance and be less representative of actual performance.

This issue of sample size bias can lead to unfair decisions—for instance, ranking a store with NPS 100 from only 3 surveys above one with NPS 80 backed by 100 surveys. To address this situation, we propose a composite metric called the Weighted Experience Index (IEP) that integrates NPS, CSAT, and the number of responses (feedback count). The goal of the IEP is to provide a more robust and fair measure of performance, reducing the impact of low response volumes and offering a more comprehensive view of customer experience at each location.

Formal Definition of the IEP

The Weighted Experience Index (IEP) is defined by a formula that combines the two experience indicators (NPS and CSAT), normalized, weighted, and adjusted by the volume of responses. Formally:

IEP = ((NPS_norm * 0.4) + (CSAT_norm * 0.3)) * (1 + log10(Feedback Count + 1) * 0.3)

Each term is defined as follows:

• NPS_norm: The normalized NPS score on a 0 to 1 scale. Since the original NPS ranges from -100 to +100, we normalize it as NPS_norm = (NPS + 100) / 200. Thus, -100 becomes 0.0 and +100 becomes 1.0.

• CSAT_norm: The CSAT score normalized to a 0–1 scale. If CSAT is given as a percentage (0 to 100%), a simple normalization is CSAT_norm = CSAT / 100.

• Feedback Count: The number of responses or surveys received for the evaluated store. In the formula, it appears inside a base-10 logarithm and is multiplied by a weight constant.

The formula first calculates a base score = (0.4 NPS_norm) + (0.3 CSAT_norm). This weighted linear combination represents customer experience performance considering both loyalty (NPS) and satisfaction (CSAT). Then it multiplies that score by an adjustment factor = (1 + 0.3 log10(Feedback Count + 1)), which incorporates the effect of response volume. Essentially, the term 1 + 0.3 log10(...) acts as a multiplier that increases (or keeps unchanged) the base score depending on how many responses were collected, giving extra weight to data richness.

Statistical Justification for the IEP Components

Each component of the IEP has been designed with statistical foundations to ensure a balanced and meaningful metric:

• Differentiated weights (40% NPS, 30% CSAT): A slightly higher weight is assigned to NPS (0.4) than to CSAT (0.3) because these metrics, while complementary, are not equivalent. NPS is designed to measure customer loyalty and overall experience (likelihood of recommendation), serving as a strong indicator of future growth. CSAT, in contrast, captures immediate satisfaction with a specific product or service interaction. Giving greater weight to NPS acknowledges its long-term strategic value, while CSAT still contributes substantially, balancing the index with short-term insight.

• Normalization of scales: Before combining NPS and CSAT, it is essential to normalize them, since they originally exist on different scales (NPS from -100 to 100, CSAT typically from 0 to 100%). Normalizing to [0,1] ensures that neither metric dominates the other just by numeric range. This also facilitates interpretation of the IEP, as the base score can be considered a proportion of maximum possible performance.

• Logarithmic volume factor: Including the number of responses addresses the issue of reliability with small samples. The IEP uses the term 1 + 0.3 * log10(Feedback Count + 1) for this. Why this form? The base-10 logarithm provides diminishing returns: the first additional responses give a relatively large improvement, but beyond a point, each new response has a smaller effect. This is desirable because increasing from, say, 5 to 15 responses should increase confidence much more than from 105 to 115. Adding 1 inside the logarithm avoids mathematical undefined values and ensures that with 0 responses, the factor equals log10(1) = 0 (i.e., a multiplier of 1.0, leaving the base unchanged).

The coefficient 0.3 acts as a volume weight: it determines how much influence volume has in the index. With 0.3, we're giving the data volume a maximum influence roughly equivalent to other metrics (approximately 30%).

Advantages of the IEP Compared to NPS or CSAT Alone

The IEP offers several advantages over using NPS or CSAT alone, especially in low-response contexts:

• Statistical robustness and reduced volatility: By incorporating response volume, the IEP reduces the impact of outliers from small samples. Thus, a store won't top rankings just because of an extreme NPS from very few surveys. Higher response counts lead to more stable, reliable scores.

• Comprehensive experience view: NPS and CSAT measure different yet complementary aspects (loyalty vs. immediate satisfaction). IEP combines both for a more holistic performance metric.

• Fair comparisons: The IEP allows more just comparisons between large and small locations, or across periods with different response rates. A branch with NPS 80 and 100 surveys can outperform one with NPS 90 and only 5 surveys, reflecting a more data-backed performance.

• Incentivizes collecting more feedback: Since more responses improve the IEP (assuming similar ratings), the metric encourages efforts to increase survey participation. This leads to richer data and better diagnostics.

Practical Examples

Example 1: High feedback volume

Store A

• NPS: 75 → NPS_norm = (75 + 100)/200 = 0.875

• CSAT: 80 → CSAT_norm = 0.80

• Feedback Count: 100 → log10(101) ≈ 2.004

IEP = (0.875 0.4 + 0.80 0.3) (1 + 0.3 2.004) = (0.35 + 0.24) (1 + 0.601) = 0.59 1.601 ≈ 0.945 (or 94.5%)

Example 2: Low feedback volume

Store B

• NPS: 90 → NPS_norm = 0.95

• CSAT: 95 → CSAT_norm = 0.95

• Feedback Count: 10 → log10(11) ≈ 1.041

IEP = (0.95 0.4 + 0.95 0.3) (1 + 0.3 1.041) = (0.38 + 0.285) (1 + 0.312) = 0.665 1.312 ≈ 0.873 (or 87.3%)

Despite having higher raw scores, Store B ends up with a lower IEP due to the low number of responses, showing the IEP’s ability to adjust for statistical confidence.

Final Reflection: Business and Academic Applications

The Weighted Experience Index (IEP) is a valuable tool for both business professionals and academic researchers seeking rigorous measurement of customer experience. In business contexts, the IEP can be embedded in CX dashboards to fairly compare branches, channels, or periods. Executives gain a unified view of service quality, knowing that a high IEP reflects not just satisfied customers but statistically solid backing.

In academia, the IEP offers an example of a statistically justified composite metric. It can be used in comparative studies of customer experience, controlling for sample size effects, or in marketing research to correlate global experience with loyalty, sales, or reputation. The index also serves as a teaching tool for indicator construction, showing how to combine quality and quantity dimensions.

In conclusion, the IEP addresses the bias inherent in traditional metrics under low-response conditions. By integrating NPS, CSAT, and response volume, it yields a more robust, fair, and actionable measure of customer experience. It empowers better business decisions and supports methodologically sound academic analysis alike.