# Losing & Regaining Points: Rating (Part 2) - Maintaining / Gaining Points

First of all, I would like to apologize for posting late. I had hoped to publish this thread before SLD but it was not until yesterday that I had some free time to work on it. Please get yourself a cup of coffee / tea or whatever you like best before starting to read this post, this time it will be a little long.

The first topic of this series was about how to calculate the amount of x* rated orders that would impact the rating causing it to lose points, this one is its complement and like the other, itâ€™s not aimed to tell sellers what to do; itâ€™s just intended to be an easy and understandable explanation on how to calculate amount of orders needed.

Structured slightly different, quick calculation tables will help this time to calculate the amount of x* rated orders for maintaining / gaining rating but, as opposed to the first topic where the given values were for not falling beyond 0.1 points, the ones given here reflect the amount of orders required for maintaining the actual rating and/or gaining points.

As we all know, rating is cumulative, it doesnâ€™t disappear from sellersâ€™ history - remaining through life for calculation purposes - and since it is an average value, it is always calculated from the very first rated order up to the moment of its evaluation (rated orders deleted by Fiverr, obviously donâ€™t count). Therefore, whatever rating shown in the profile, it is the overall rating or as Fiverr calls it, â€śAll-Time Ratingâ€ť.

It should also be noted that rating shown by Fiverr is rounded to 1 decimal digit, which means that every number within a 0.099 interval will be expressed as X.Y (e.g., [4.855 - 4.954] will be rounded to 4.9) - rounding, by the way, that doesnâ€™t tell much about how close anyone is to either edge. This point seems just as a math explanation but itâ€™s not if Fiverr takes into account this range of values when calculating the rating, reason why â€śexpanded rating tablesâ€ť will be incorporated this time.

Likewise, it is important to mention the following about the amount of orders (AO) needed to modify the rating - whether positively or negatively:

• AO is directly proportional to the total amount of reviews, meaning that the greater the amount of reviews, the more new rated orders will be needed to either to fall or jump. In this regard, changes in rating are less drastic as the total amount of reviews increases (See graphs 1 & 2).

• AO is an asymptotic function, making it easier or harder to fall, keep or raise the rating depending on: 1) the proximity of values between the rate of the new orders and the rating to be kept/achieved and 2) the location of the actual rating within the interval - on either edge of any given rating interval, AO tend to behave in a similar way regardless of the amount of reviews (See graphs 3 through 5).

Throughout the graphs, the following statements can be verified:

• Less amount of rated orders, bad/good respectively, is requiered for maintaining rating than for gaining it.

• Changes in rating are more drastic as the total amount of reviews decreases, thereby, sellers with low amount of reviews can get into trouble very easily.

• Flat surfaces might not be as wide as one may think when looking to the 1 decimal digit rating showing in each oneâ€™s profile, since the nearer the rating is to the lower edge, the lesser amount of orders it will take to fall. Same happens for gaining points when rating is nearer to the upper edge - lesser amount of orders will be needed to jump.

• Amount of rated orders tend to behave in a similar way regardless of the amount of reviews when any given rating is on either edge of the interval.

• Once fallen, the easiest way to recover, is by trying to get orders to be rated as far as possible from the actual rating not only because the rate would be higher, but because of being further away from the rating that wants to be raised - the nearer the rate is from the rating, the higher the amount of orders of that specific rate will be needed (asymptotic approximation).

Additionally, for the same reason of being asymptotic, if anyone thinks that jumping straight into 0.2 points is just multiplying by 2 the amount of orders that makes the rating raise 0.1 points, bad news, itâ€™s not. So the best way to climb, is one step at a time and as explained above.

Considering what has been said and the fact that I donâ€™t know Fiverrâ€™s criteria nor the amount of decimal digits involved in calculations, the methodology used will differ this time because it is much more important for values to be as accurate as possible. Likewise, formulas will also be provided.

Things to be taken into account:

• Values on the quick calculation table are based assuming Fiverr rounds up with 0.055 (e.g, 4.455 = 4.5) and, although combination of rated orders can also be made, each value was determined taking into consideration only the specific rate as if it were the only one to impact the overall rating.

• Values on the expanded rating tables are calculated on a 0.001 step basis and for correcting purposes assuming Fiverr calculates rating with 3 decimal digits before rounding to 1 decimal digit.

• Divided into 2 areas, the quick calculation table will serve this time on double purpose: yellow is for maintaining rating and blue is for gaining rating. All values work and can be applied directly for any given amount of reviews, even with small amounts.

• Values shown in the quick calculation table are given as a percentage of the total amount of reviews and values shown in the expanded rating tables are just factors by which to multiply depending on the 3 decimal digits rating.

Calculating Actual Rating:

Go to the bottom of your â€śAnalyticsâ€ť page were it says â€śRatingsâ€ť and multiply each number that appears in parentheses with its corresponding star. Add all the products and divide by the total amount of reviews. Leave the obtained number with the first 3 decimal digits (e.g., 4.464546 = 4.464). That will be your expanded rating.

Finding the amount of rated orders: Using the tables - Methodology

A.- Multiply your total amount of reviews by the corresponding value shown on the quick calculation table.

• Position yourself on the row that corresponds to the 1 digit rating, the whole row will give you the complete range of values: yellow for maintaining and blue for gaining. Intersect with the column corresponding to the rate youâ€™re looking for and multiply the intersection value by the amount of reviews.

B.- Multiply the obtained number by the factor shown on the corresponding expanded rating table and divide by 100.

• Position yourself on the row that corresponds to the 2nd decimal digit rating and intersect with the column corresponding to the last decimal digit rating. Multiply the intersection value (factor) that corresponds, whether for maintaing or gaining, by the number obtained in A and divide by 100.

For Maintaining Rating: Only use the integer part of the number obtained in B (e.g., 49.46 = 49 ; 52.87 = 52).

For Gaining Rating: ADD 1 to the number obtained in B and only use the integer part (e.g., 49.46 = 49 + 1 ; 52.87 = 52 + 1).

Example:

Letâ€™s say you have a 4.4 rating and want to know how many 4.8* orders will be needed to pull it up and what would be the maximum amount of 3.5* orders that can be afforded for not dropping.

Rating shown on profile: 4.4*
Reviews: 58 reviews

1.- Calculate your actual rating to 3 decimal digits

Detailed Rating showing on the Analytics page:
5 (37)
4.7 (9)
3.7 (5)
1.7 (2)
1 (5)

((5 * 37) + (9 * 4.4) + (6 * 3.7) + (1.5 * 2) + (1 * 4)) / (37 + 9 + 6 + 2 + 4)

(185 + 42.3 + 18.5 + 3.4 + 5) / 58

254.2 / 58 = 4.38275â€¦

Expanded rating = 4.382

2.- Use the quick calculation table & expanded rating tables

• Maintaining rating: Using the yellow area

A.- Position yourself on the row that corresponds to 4.4 on the quick calculation table, intersect with the 3.5 column and multiply the intersection value by the amount of reviews.

(58 * 5.26) = 305.08

B.- Position yourself on the row that corresponds to 0.080 on the expanded rating table for maintaining rating and intersect with the 0.002 column. Multiply the factor with the value obtained previously and divide by 100. Keep the product and just use the integer part.

(305.08 * 0.602) / 100 = 1.836

Maximum amount of 3.5 * orders for maintaining rating = 1

• Gaining rating: Using the blue area

A.- Position yourself on the row that corresponds to 4.4 on the quick calculation table, intersect with the 4.8 column and multiply the intersection value by the amount of reviews.

(58 * 15.94) = 924.52

B.- Position yourself on the row that corresponds to 0.080 on the expanded rating table for gaining rating and intersect with the 0.002 column. Multiply the factor with the value obtained previously and divide by 100. Add 1 to the product and just use the integer part.

(924.52 * 1.326) / 100 = 12.259

12 + 1 = 13

Minimum amount of 4.8 * orders for gaining raiting = 13

P.S.: Next post will be about Order Completion Rate (OCR)

21 Likes

Wow, youâ€™re a mathematics geniusâ€¦I love how you always used mathematics to explain your key pointsâ€¦

2 Likes

Hi @wordpress_whiz, thanks once again for your nice words! You are very kind!

As I said before, numbers are easier for me to speak through. They are fun, easy and quick, and they also help me to relax a bit putting my mind away from what Iâ€™m currently doing.

As always, the difficult part is writing down the post

3 Likes

Hello, thatâ€™s very impressive. I donâ€™t want to focus on ratings too much since I value maintaining my peace of mind.

As long as there is a part of the ratings process I have zero control over why worry about it? I just do my best at all times.

3 Likes

Hi @misscrystal,

These posts are intended to teach sellers on how to obtain numbers and with them, planify and make estimations on whatever they need.

Forum is full of people asking how x* orders and their amount (specially 1* orders), affect their rating and how to calculate it. There are also users that ask how many 5* orders they need to raise their rating. Well, now they can learn how and take action at any time before it is too late or without waiting to things happen.

Any user can download the tables and the formulas to use for their own advantage and without needing to overflow the forum asking for advice in what to do. Now they are able to take their own decision on whether to cancel or to take the possible bad review because now they know the how and the how many that will affect them.

Knowledge gives people control and thatâ€™s good.

Next post: Order Completion Rate (OCR) - Cancel or not cancel, that is the question

6 Likes

Impressive stuff @maitasun

Would drive me insane applying this to my orders/ratings but sure itâ€™ll be handy to folk

2 Likes

This is impressive!!!

2 Likes

Hi @smork,

Itâ€™s not really that difficult, just needs some practice and thatâ€™s it! Although you donâ€™t need it and have nothing to fear for, your amount of reviews is very high (3k+) and your rating is excellent (5*)!

Yes, youâ€™re right and thatâ€™s what I hope for. One of the reasons for teaching sellers, is making them independent and capable of deciding by themselves what to do in their best interest.

Rating is affected differently depending on who. Regarding amount of reviews and orders, there are 4 categories of sellers:

• Sellers with big amount of reviews and high volume of orders.
• Sellers with great amount of reviews and low volume of orders.
• Sellers with low amount of reviews and big volume of orders (newly started).
• Sellers with low amount of reviews and low volume of orders.

For the first 2 type of sellers (which is your case as well as many other old sellers), losing rating and therefore, having to gain it back, is not an issue unless they start to continously deliver late or very poor quality job, etc.

As explained, rating is an average value of the total amount of reviews, meaning itâ€™s volume based. So, the greater the total amount of reviews, the lesser the rating changes. It would be needed a big amount of bad rated orders in a given period to have an impact in the rating in that same time span.

Think of it as your college score: when youâ€™re new, bad scores impact heavily on your general score but as you stand along time, raising or lowering the general score is more difficult as it doesnâ€™t vary that much.

For instance, someone with 3k reviews and 5* rating would have to get 46 orders rated 2* to drop to 4.9*, but for someone with 100 reviews and 5* rating, only 2 orders rated 2* will make the rating fall to 4.9* and, depending on how Fiverr calculates rating, it will take anywhere from 32 to 123 orders rated 5* to bring the rating back to 5*.

As you see, the ones to be affected by bad rated orders are those with low amount of reviews, specially, those with low amount of income orders (low volume of orders) because once they fall, theyâ€™ll need many orders to gain the lost points.

1 Like

Ohhhhh!!! I didnâ€™t notice up to now that the post was pinned!!!

Thank you!

1 Like

Iâ€™m going to repeat my infamous line - Iâ€™m a writer, not a mathematician.

Awesome job explaining it.

1 Like

Thanks Emerald, I do appreciate your words!

For me is the other way around, I donâ€™t belong to the world of words but of numbers. Being so, Iâ€™m glad once again I was able to write down my ideas in a coherent way so to be as helpful as possible and that you all find my explanation understandable.

2 Likes

â€¦ Wow!

These 2 points explain why long-established sellers donâ€™t have to worry anymore for getting low-rated orders and refusing request for cancellation, meanwhile most new sellers opt for order cancellation when encountering hassle instead of getting low ratings.

By the way, i never imagine that the rating behaviour curves are so steep.

You did a great job on explaining the whole thing and making the visuals!

2 Likes

Hi @pandumstudio, thanks!

Yes, the best way to go for those with high amount of reviews (normally old sellers) is refusing to cancel while for new/unleved sellers is the opposite, i.g., cancelling to avoid a poor review.

Leaving aside some circumstances, leveled sellers have high amount of reviews to absorb low/bad ones making it difficult to negatively affect their rating. Instead, they need to be worried about cancelations as these would put them in trouble due to demotion.

New/unleveld sellers with low amount of reviews, on the other hand, donâ€™t have to worry about demotion, as they donâ€™t have a level to risk - at most, they only risk promotion - but cannot afford bad reviews because they donâ€™t have enough amount of good ones to absorb the negative impact. Thatâ€™s why the best way for these sellers is to cancel.

Now, thereâ€™s a special group of sellers - the ones in the middle of the sandwich (with not too high and not too low amount of reviews) - where they have to ponder what hurts the less: a cancellation or a bad review. That decison will only depend on what they are willing to risk, whether rating or level.

Yes and the lesser the amount of reviews, the steeper the rating is - either for falling or gaining. Thatâ€™s the reason why low amount of reviews can lose or gain many points with only a few orders.

4 Likes

You are genius
Great post

1 Like

good job, with detailed information to keep a rating control.
Thank you for the time given in carrying out this work and giving us the information.

1 Like

Thanks @guilarteluis

Thatâ€™s the idea, to give sellers knowledge - which means power - and tools for them to have control and decide for themselves.

Hereâ€™s the other topic (last of the series) you may wnat to take a look at, hope you find it useful as well