Interpreting customer service metric ratings

This topic describes how to understand the information returned from a call to getCustomerServiceMetric.


Getting customer service metric ratings

You can configure getCustomerServiceMetric to return one of four different customer service metric ratings.

A request can return either the CURRENT or PROJECTED seller rating for either the Item Not As Described or Item Not Received customer service metric.

Configure a combination of these two sets of query parameters in your call's request to get the benchmark data and metric rating of your choosing.

Important! Sellers are eligible to be rated only if they have sold an item in the last 12 months. In addition, they can receive a CURRENT rating only if they sold an item during the last lookback period.

Calculating customer service metric ratings

Customer service metric ratings are based on how a seller performs over a set period of time compared to the benchmark that is set over that same period of time.

Note: The benchmark data, seller data, and seller ratings returned by getCustomerServiceMetric are all based on the parameters established in the returned evaluationCycle container and the metric data is confined by the values returned in the dimensionMetrics.dimension container.

To simplify descriptions, these constraints are assumed and are omitted in the following discussions.

Benchmark data and seller ratings

Each customer service metric rating is based on how the seller compares to the benchmark data collected for that same metric over the same period of time.

Sellers are rated on a scale and benchmark.rating returns a value from LOW to VERY_HIGH, depending on the deviance of the seller's customer service metric value to the benchmark average. Here, a LOW deviance is best and a VERY_HIGH rating reflects poorly on the seller.

Currently, only the PEER_BENCHMARK is considered, where the seller is evaluated against benchmark data collected on their peer sellers. Peer sellers is the group of sellers that offers similar products under similar selling circumstances (including selling prices, and terms of sale and shipping destination).

The benchmark average rating is set to the average value of the collected benchmark data for the metric. The seller's rating is determined by comparing the seller's performance value to the average value of the collected benchmark data.

The following example displays the response data for a LOW customer service metric rating:

    "metricKey": "RATE",
    "value": "0.47",
    "benchmark": {
      "basis": "PEER_BENCHMARK",
      "rating": "LOW",
      "metadata": {
        "average": "0.48"

Note that if the value of the metrics.value field falls below a certain threshold, the benchmark.rating field returns NOT_APPLICABLE, meaning the seller did not have enough transactions for eBay to calculate a customer service metric rating.

Lookback periods

The period over which a seller (and the associated benchmark data) is evaluated is known as the "lookback period."

A lookback period can either be 3 months or 12 months, depending on transaction volume of the seller.

Lookback periods always start on the first day of a month (specifically at midnight, Mountain Standard Time) and they always end on the last day of a month at (11:59.59 PM, also MST).

Seller data for the lookback period is evaluated on the 20th of the month following the conclusion of the lookback period, at midnight, Mountain Standard Time. Note that the CURRENT seller rating is not updated until the evaluation occurs for the lookback period.

For example, suppose a seller has enough volume to be rated on a three-month lookback period, and they review their CURRENT customer service metric rating on October 21st:

  1. The evaluation would have occurred the previous day (October 20th)
  2. Their three-month lookback period would have started on July 1st (at midnight, MST) and would have ended on September 31st (a 11:59.59 PM, MST).

The response, time stamps returned in the evaluationCycle container are formatted as ISO 8601 strings based on the 24-hour Coordinated Universal Time (UTC) clock. Coordinated Universal Time (UTC) is 6 hours ahead of Mountain Standard Time (MST), which is reflected in the timestamps returned.

For example:

  "evaluationCycle": {
    "evaluationType": "CURRENT",
	"evaluationDate": "2019-10-21T19:13:00.000Z",
	"startDate": "2019-07-01T07:00:00.000Z",
	"endDate": "2019-10-01T06:59:59.000Z"

Current versus projected ratings

Sellers can request either a CURRENT or PROJECTED customer service metric rating.

The current seller rating is the rating that was established for the last lookback period evaluated. With normal seller activity, a new CURRENT customer service metric rating is established each month on the 20th.

The projected seller rating is a daily evaluation that provides a running projection of how the seller is performing within the lookback period currently being evaluated. In other words, this is a non-certified metric rating for how the seller is doing since their last evaluation.

Metric keys

In the response, each set of customer service metrics is divided into three metric keys:

  • RATE – When metricKey is set to RATE, the fields in the associated container return the seller's calculated value for the associated customer service metric along with the benchmark data against which the seller is evaluated. The benchmark.rating value is the seller's rating for the metric in the associated dimension and evaluationCycle.
  • COUNT – When metricKey is set to COUNT, the associated value field returns the number of transactions completed by the seller for the metric being evaluated in the associated dimension and evaluationCycle.
  • TRANSACTION_COUNT – When metricKey is set to TRANSACTION_COUNT, the associated value field returns the number of transactions completed in the peer group for the metric being evaluated in the associated dimension and evaluationCycle.

Customer service metric types

Sellers can retrieve their customer service metric rating for two different metrics. The metrics are defined by the following transaction types:

  • ITEM_NOT_AS_DESCRIBED (abbreviated as INAD)
  • ITEM_NOT_RECEIVED (abbreviated as INR)

Set the metric rating you want returned using a query parameter in your call to getCustomerServiceMetric.