Jump to content

Suggestion: Change how Average Response Time is shown/calculated.


uk1000

Recommended Posts

Let's say there's a professional business with a website and they miss responding to a message for a couple of days or even a week or more (and that business might only work Monday to Friday, so unlike Fiverr sellers they might not be expected to respond on some days).

So if that business had missed responding to a message for some time, the professional thing to do would be for them to respond to it, maybe apologising if there was a long response time.

But Fiverr's system discourages sellers from doing the professional thing and responding to a message they missed for a day or more.

eg. If a seller responds to the only message they have received in the last 90 days and it was done 3 days after the buyer's message was sent. If the seller does that, their profile, seen to all buyers, will show "Avg Response Time: 72 hours". And it will show that for nearly 90 days, so buyers will see it and probably decide against buying.

It will look even worse if they miss a message for about week , and the seller does the professional thing and responds to that message (and it's the only message the response time calc takes into account). It will then show on their profile to all buyers something like "Average response time: 168 hours". In the old profile format it was at least shown in days (when it was >=about a day) where it didn't look as bad.

So if the seller does the professional thing and respond to the old message, they will likely get the really high avg response time shown on their profile, maybe for nearly 3 months, discouraging other buyers from ordering. Or they can do the thing that won't be that professional by not responding at all to that message and keeping the Fiverr average response time not showing or not showing a very high figure (though if it's been very long the buyer might have found someone else anyway).

Shouldn't certain responses be excluded from the avg response time calculation (with it giving info on that in a help part of the profile page, seen by buyers?), so that sellers could respond to messages that are already more than 24 hours late without further penalty than the already lowered response rate %? Maybe allow sellers to give/select explanations for very late responses. Or at least it should show the avg response time in days (or other format) rather than hours when it's over 24 hours. Maybe there's a seller who only works some days of the week (eg. Monday to Friday, like other businesses), shouldn't the system allow the response time calc to only take into account office hours of the seller?

Edited by uk1000
  • Like 8
Link to comment
Share on other sites

I have noticed the exact same thing, and have given feedback to Fiverr support and to my success manager regarding how Fiverr calculates message response rate as a percentage, and how this leads to an exaggerated negative impact on the account during times of low traffic. What you've seen here is the tip of the iceberg.

A single late response during times of high volume traffic and enquiries to your gig will have a negligible impact on the account, but a single missed message during a period of low traffic and enquiries will have a devastating impact - because it will lead to reduced traffic itself, and the account will enter a cycle of increasing penalties entirely beyond the seller's control.

This occurs as a result of calculating response rate as a percentage (over a fixed period of time), which is the least sophisticated metric you can use, but the easiest to plug in to an algorithm.

The success scores are calculated using the same kind of basic and overly simplistic metrics. This is why they do not nor will they ever reflect reality except (purely by coincidence) for a small chunk of sellers.

There is already a penalty for late message response - the seller will lose the work. No other artificial penalty is necessary nor should be desired by Fiverr. A smarter (and more accurate) metric for conversion rate can be calculated from messages received -> order placed. At the moment this is calculated from clicks, which again is an unsophisticated and basic metric which is coming from the tech world/website/online sales world.

Once you dive into how these metrics are calculated (thanks to the seller success scores now being shown to us) it becomes evident that the entirety of the algorithm that drives site rankings is fed by poorly designed metrics and in some cases simply by 'bad data'.

An example of 'bad data' is the private feedback mechanism driving the customer satisfaction rating on a gig. Fiverr have chosen a fixed time period from which to draw their data - but one seller's gig might have 500 pieces of private feedback and another seller may only have 5 pieces of private feedback for the same time period. So again we see the same issue as in the message response time - one piece of negative feedback in a small sample has an exaggerated negative impact, and one negative feedback in a larger sample has a negligible impact. Both sellers have the same explicit amount of negative feedback but will have wildly differing customer satisfaction scores. As with the message response rate, there will be a chunk of sellers for whom this will reflect reality, and if you were to look closely you'd see that they would all have roughly the same amount of orders/feedback for the same time period.

It is entirely beyond me that no one has yet been able to articulate the obvious and easily demonstrable deep, deep flaws with the metric calculations on Fiverr. You could write a paper on it.

I have provided written feedback to Fiverr on all of the above and I have not received any indication that they actually understand the issue with their metrics, which is troubling.

 

I hope that the above information is useful and informative. To briefly summarise it - during times of low traffic and enquiries to your gigs, Fiverr metrics present a high risk to your account and you should take steps to manage that risk.

 

"...the professional thing to do would be for them to respond to it, maybe apologising if there was a long response time. But Fiverr's system discourages sellers from doing the professional thing and responding to a message they missed for a day or more."
- too many sticks and no carrots

  • Like 3
  • Up 1
Link to comment
Share on other sites

Hi there! 

These are some valid points and I thank you for these suggestions! We are always thrilled to hear feedback from you and others who use Fiverr. I’ll be sure to pass this along to the rest of the team for consideration as we continue to brainstorm ways to improve the Fiverr experience for all. 
 

  • Like 1
  • Thanks 2
Link to comment
Share on other sites

3 minutes ago, necramedia said:

I respond within a few minutes or at least a hour, and my response rate keeps being stuck at 83 and 84%. And if you block someone, like a scammer, fiverr states it does not affect your response rate, but how come my response rate still goes down? 

Did you reply to the scammer before blocking them? If not, then it can affect your response rate

  • Like 2
  • Up 3
Link to comment
Share on other sites

24 minutes ago, necramedia said:

like a scammer, fiverr states it does not affect your response rate, but how come my response rate still goes down

If it does then it's a bug that still exists. I thought they said it was fixed at one time but if it still exists you could report it to CS and maybe they'll try to get it fixed.

But for now you could respond and then block/report like suggested.

Edited by uk1000
  • Like 2
Link to comment
Share on other sites

24 minutes ago, dponzio said:

Isn't it ridicolous to be obliged to reply to scammers? 😉

It is 💯

But that has always been the case. I believe system can't differentiate between a normal buyer and a scammer - so every first text gets accounted for our response rate. 🫤

Edited by priyank_mod
  • Like 3
Link to comment
Share on other sites

Quote

my response rate keeps being stuck at 83 and 84%. 

I was also stuck at 83% by missing a single message during a period of low messages. Mine was stuck here until my Success Manager cleared it, support would not.

The average response time isn't used to calculate the reduction on your response rate %, what it does is it flags a message as missed after 24 hours, and the system will calculate a negative impact based on how many messages are flagged as missed in the last X months. It's straightforward to see how it's working, the only data you don't have is the time period it is using - I suspect it's the last 3 months.

 

So as I have detailed above, a single missed message during a low traffic period will put your gig into a downward spiral, because that negative impact will increase as the total messages you receive decreases.

If a lower percentage response time reduces your visibility on the platform (which I suspect it does) then naturally you will receive less messages and the response rate % will continue to drop!

One missed message during a time of low traffic will put the gig into a death spiral - because they have a metric which measures response rate as a percentage of total messages, and an algorithm that will reduce your total messages received BASED ON RESPONSE RATE.

It continues to baffle me why no one can see the obvious problem here.

It's all bad data they are working with. 

 

 

  • Like 2
Link to comment
Share on other sites

On 7/10/2024 at 6:31 PM, priyank_mod said:

It is 💯

But that has always been the case. I believe system can't differentiate between a normal buyer and a scammer - so every first text gets accounted for our response rate. 🫤

thanks for sharing this

Link to comment
Share on other sites

Please sign in to comment

You will be able to leave a comment after signing in



Sign In Now
×
×
  • Create New...