|
 |
| | | Author: | WUIt  | | Posted: | Feb 15, 2022 15:31 | | Subject: | New feedback and rating system | | Viewed: | 122 times | | Topic: | Suggestions | | Status: | Open | | Vote: | [Yes|No] | |
|
| The current BrickLink feedback system with all stores having unrealistically
high (80%) praise-ratings is inadequate from a buyers' perspective. When
evaluating stores, a buyer looks at 1) terms and conditions, 2) seller location,
3) store age, 4) number of completed transactions, 5) likelyhood to deliver all
parts, 6) quality of packaging, and 7) speed of service. The praise-percentage
rating as currently used cannot distinguish the stores delivering excellent service
from the mundane. There is no way to praise a store for good packaging, broad
sortiment, identification of different molds, fast delivery, or exceptional service.
There is also no way to express dissatisfaction in sortiment, slow service, missing
parts, poor packaging, unexpected condition/color of parts, unclear terms&conditions,
unexpected/hidden costs, or surprising shipment costs.
Non-paying buyers and non-shipping sellers have their own dedicated functions,
so these issues must not be part of the feedback system. The neutral feedback
response must assume that the buyer pays the invoice and the seller delivers
the merchandise. Any deviations (positive or negative) from this default expectation
can be considered measures of quality.
The new order feedback system is proposed to have 5 potential responses; Excellent
(Outstanding), Pleased (Good), Adequate (Neutral), Disappointed (Discontent),
and Apalled (Horrible). No feedback should imply an incomplete transaction.
Feedback older than 12 months must not affect the rating. Every Excellent order
rating must exactly counter an Apalled order rating. Every Pleased order rating
must exactly counter a Disappointed order rating. Excellent must improve rating
more than Pleased. Disappointed must reduce order rating more than Adequate.
A new store must not have an untruthful rating. A "perfect" rating should not
be achievable in practise.
Applying the standard solution (from polls and surveys) to this context, results
in the formula:
store rating = 100 * (2*E + P - D - 2*A) / (8 + 2*E + P + N + D + 2*A),
where E = number of orders marked Excellent, P = number of orders marked Pleased,
N = number of orders marked Adequate, D = number of orders marked Disappointed,
and A = number of orders marked Apalled.
The potential rating ranges from -100 to 100 with both extremes extremely difficult
to achieve. A new store starts with rating = 0 indicating that it is neither
good nor bad. If a store has constant ratio 1 of positives to negatives, then
the rating will increase slightly over time. A large number of neutrals will
push the rating back toward zero.
This proposed system would eliminate the misunderstandings regarding "praise"
in the present system, allow honest feedback, and relieve all sellers from fearing
neutral feedback or retaliation. Additional benefits include a rating able to
separate the great from the ordinary and a more accurate indicator of quality.
|
|
|
| | | | | |
| | | | | Author: | marlinjc | | Posted: | Feb 15, 2022 17:21 | | Subject: | Re: New feedback and rating system | | Viewed: | 31 times | | Topic: | Suggestions | |
|
| In Suggestions, WUIt writes:
| | The current BrickLink feedback system with all stores having unrealistically
high (80%) praise-ratings is inadequate from a buyers' perspective. When
evaluating stores, a buyer looks at 1) terms and conditions, 2) seller location,
3) store age, 4) number of completed transactions, 5) likelyhood to deliver all
parts, 6) quality of packaging, and 7) speed of service. The praise-percentage
rating as currently used cannot distinguish the stores delivering excellent service
from the mundane. There is no way to praise a store for good packaging, broad
sortiment, identification of different molds, fast delivery, or exceptional service.
There is also no way to express dissatisfaction in sortiment, slow service, missing
parts, poor packaging, unexpected condition/color of parts, unclear terms&conditions,
unexpected/hidden costs, or surprising shipment costs.
Non-paying buyers and non-shipping sellers have their own dedicated functions,
so these issues must not be part of the feedback system. The neutral feedback
response must assume that the buyer pays the invoice and the seller delivers
the merchandise. Any deviations (positive or negative) from this default expectation
can be considered measures of quality.
The new order feedback system is proposed to have 5 potential responses; Excellent
(Outstanding), Pleased (Good), Adequate (Neutral), Disappointed (Discontent),
and Apalled (Horrible). No feedback should imply an incomplete transaction.
Feedback older than 12 months must not affect the rating. Every Excellent order
rating must exactly counter an Apalled order rating. Every Pleased order rating
must exactly counter a Disappointed order rating. Excellent must improve rating
more than Pleased. Disappointed must reduce order rating more than Adequate.
A new store must not have an untruthful rating. A "perfect" rating should not
be achievable in practise.
Applying the standard solution (from polls and surveys) to this context, results
in the formula:
store rating = 100 * (2*E + P - D - 2*A) / (8 + 2*E + P + N + D + 2*A),
where E = number of orders marked Excellent, P = number of orders marked Pleased,
N = number of orders marked Adequate, D = number of orders marked Disappointed,
and A = number of orders marked Apalled.
The potential rating ranges from -100 to 100 with both extremes extremely difficult
to achieve. A new store starts with rating = 0 indicating that it is neither
good nor bad. If a store has constant ratio 1 of positives to negatives, then
the rating will increase slightly over time. A large number of neutrals will
push the rating back toward zero.
This proposed system would eliminate the misunderstandings regarding "praise"
in the present system, allow honest feedback, and relieve all sellers from fearing
neutral feedback or retaliation. Additional benefits include a rating able to
separate the great from the ordinary and a more accurate indicator of quality.
|
The proposal is too complicated. As is, I only get feedback on about 60% of
my sales. Many do not even indicate received or completed. My feedback responders
often comments on packaging,promptness etc.
|
|
|
| | | | | |
| | | | | Author: | marlinjc | | Posted: | Feb 15, 2022 17:22 | | Subject: | Re: New feedback and rating system | | Viewed: | 38 times | | Topic: | Suggestions | |
|
| In Suggestions, WUIt writes:
| | The current BrickLink feedback system with all stores having unrealistically
high (80%) praise-ratings is inadequate from a buyers' perspective. When
evaluating stores, a buyer looks at 1) terms and conditions, 2) seller location,
3) store age, 4) number of completed transactions, 5) likelyhood to deliver all
parts, 6) quality of packaging, and 7) speed of service. The praise-percentage
rating as currently used cannot distinguish the stores delivering excellent service
from the mundane. There is no way to praise a store for good packaging, broad
sortiment, identification of different molds, fast delivery, or exceptional service.
There is also no way to express dissatisfaction in sortiment, slow service, missing
parts, poor packaging, unexpected condition/color of parts, unclear terms&conditions,
unexpected/hidden costs, or surprising shipment costs.
Non-paying buyers and non-shipping sellers have their own dedicated functions,
so these issues must not be part of the feedback system. The neutral feedback
response must assume that the buyer pays the invoice and the seller delivers
the merchandise. Any deviations (positive or negative) from this default expectation
can be considered measures of quality.
The new order feedback system is proposed to have 5 potential responses; Excellent
(Outstanding), Pleased (Good), Adequate (Neutral), Disappointed (Discontent),
and Apalled (Horrible). No feedback should imply an incomplete transaction.
Feedback older than 12 months must not affect the rating. Every Excellent order
rating must exactly counter an Apalled order rating. Every Pleased order rating
must exactly counter a Disappointed order rating. Excellent must improve rating
more than Pleased. Disappointed must reduce order rating more than Adequate.
A new store must not have an untruthful rating. A "perfect" rating should not
be achievable in practise.
Applying the standard solution (from polls and surveys) to this context, results
in the formula:
store rating = 100 * (2*E + P - D - 2*A) / (8 + 2*E + P + N + D + 2*A),
where E = number of orders marked Excellent, P = number of orders marked Pleased,
N = number of orders marked Adequate, D = number of orders marked Disappointed,
and A = number of orders marked Apalled.
The potential rating ranges from -100 to 100 with both extremes extremely difficult
to achieve. A new store starts with rating = 0 indicating that it is neither
good nor bad. If a store has constant ratio 1 of positives to negatives, then
the rating will increase slightly over time. A large number of neutrals will
push the rating back toward zero.
This proposed system would eliminate the misunderstandings regarding "praise"
in the present system, allow honest feedback, and relieve all sellers from fearing
neutral feedback or retaliation. Additional benefits include a rating able to
separate the great from the ordinary and a more accurate indicator of quality.
|
The proposal is too complicated. As is, I only get feedback on about 60% of
my sales. Many do not even indicate received or completed. My feedback responders
often comments on packaging,promptness etc.
|
|
|
|
|
|
Finland, Uusimaa
USA, Illinois