Kindle Forum banner
1 - 20 of 99 Posts

·
Registered
Joined
·
4,279 Posts
Discussion Starter · #1 ·
Need moar data!

Sooo, it's beginning to look like a long-lasting glitch or a MAJOR algo change in the freebie ranks hit on Sept 15.

Let's figure out which! I'm collecting data from a pool of 150 freebie owners today, although not everyone there qualifies to play or will choose to play. You can help add to the collective IF your book meets the following criteria:

A) It's on Select free promo right now
OR
B) It's a permafree on promo right now

AND (applies to BOTH A and B)

C) It has MORE sales today than it did yesterday.

Got that? (A or B) + C

We need to be consistent to get an accurate picture of what's going on. If we're watching rank, that means we need books that are gaining to get to that rank, not just passing through it as they fall or maintaining it because of bloated history.

If your title meets the requirement, here's what I would love for you to do:

* In the US morning after about 5am and before 10am CT (6-11am ET, 3-8am PT), jot down the book's rank on .com.

* Check your Sales Graph and see how many downloads you had today (Wednesday) ONLY on the .com site. That's very important to restrict it to .com.

* Record the stats here (or PM me):

- Your book's ASIN* (in case I have questions, so I can check the Tracker tool or the product page, etc) - I won't share the ASIN directly beyond this forum or the PM. You can also provide downloads and rank here without the ASIN to maintain your anonymity. If it fits the data curve, I'll love to include it. If it doesn't fit, and I can't verify what and why, I'll have to ignore it.

- # of downloads on .com only
- Current rank on .com
- Select of Permafree


Example:

B008THBM5S
277 DLs
#1,740
Select
Let's see how tomorrow goes and we'll probably do it on Friday morning too. A book can play just once on either day or on both days, providing it continues to meet eligibility requirements.

I'll post out a chart (without the ASINs) when we're all done, compare it to what life was like pre-Sept 15, and together we'll see if we can make sense of it.

Oh, and if you're in the SPRT FB group, too, please just post here OR there.

Thank you!
 

·
Registered
Joined
·
274 Posts
PhoenixS said:
* In the US morning after about 5am and before 10am CT (6-11am ET, 3-8am PT), jot down the book's rank on .com.

* Check your Sales Graph and see how many downloads you had today (Wednesday) ONLY on the .com site. That's very important to restrict it to .com.
Just to clarify, we're checking the sales rank tomorrow (Thursday) at (for me) 3-8AM PST? And recording only Wednesday's downloads?

I have a Bookbub on a Select freebie today so should be able to report if I can remember to get on the computer that early. :p

P.S. I would love to hear more about this glitch. What do you suspect? *gazes with rapt curiosity at Phoenix*
 

·
Registered
Joined
·
4,279 Posts
Discussion Starter · #3 ·
LindsayBuroker said:
Just to clarify, we're checking the sales rank tomorrow (Thursday) at (for me) 3-8AM PST? And recording only Wednesday's downloads?

I have a Bookbub on a Select freebie today so should be able to report if I can remember to get on the computer that early. :p

P.S. I would love to hear more about this glitch. What do you suspect? *gazes with rapt curiosity at Phoenix*
Pfft. :-*

Correct on the procedure., Lindsay. Although BookBub'd books are probably not the best candidates for this. Massive enough downloads are breaking through whatever the issue is. So BB books are still hitting in the Top 10. What is impossible to know when there is that much competition for the top spots is what and where the true hit is.

For instance, a couple of weeks ago, we had a BB book on Day 2 of being Select free. On Day 1, the book had 1028 .com downloads. Best rank until the BookBub mail went out was #1594. Obviously this book was affected. It had 32,200 .com DLs on Day 2. It hit a rank of #4. Were the previous 1000 downloads still not being seen by the algo? If they HAD been seen, would the rank have been #3? Well, at that level, it's impossible to know because we don't know the circumstances of Book #2 or #3. Like with the Paid rank, once you're in to the Top 10, it's pretty much a free-for-all. #5 could have 5000 sales, while #4 could have 10,000. Between #20 and 100, the list is MUCH less volatile and much more predictable.

Do share your numbers, because all data is instructive! I just don't think it will be AS instructive in this case.

I only have limited data right now, which is why I need MOAR.

<speculation>
Speculation ONLY, it seems what I'm desperately hoping is a glitch cropped up on Sept 15. It SEEMS to affect only newly promoted titles -- permafrees on promotion and Select titles going free. Permafrees maintaining their rank ranges seem to maybe still be properly inserted into the calculations. Titles that should easily be in the Top 100 or in any rank range, really (other than the Top 20 tier) aren't coming close to where they would have ranked pre-Sept 15.

Great if this were a way to punish scammers and suppress their influence, but the very titles that ARE getting through seem to only be the titles with massive downloads. So this is actually HELPING the scammers who work with bots grab and hold visibility. So either it was an algo change gone bad OR a glitch. But again, I'm working with limited data right now -- 11 titles that I manage, half of which had no promo mention anywhere during their free days so my feeling is that it's not an impact against any one ad site or group of ad sites, and another 11 in a current group promo, all of which experienced the glitch yesterday whether they had ad support or not (the group promo itself begins today).</speculation>

So waiting to add more data points before I call KDP and try to find someone who knows what I'm even talking about to explain the issue to. Or to regroup if this is the new reality. *shudder for anyone unable to hitch a ride on BookBub or the BotMobile.*
 

·
Registered
Joined
·
1,125 Posts
PhoenixS said:
So this is actually HELPING the scammers who work with bots grab and hold visibility.
And pointing out to Amazon which eBooks are riding the BotMobile (I'm totally stealing this word!) if they also follow certain other patterns.
 

·
Registered
Joined
·
4,279 Posts
Discussion Starter · #5 ·
Gator said:
And pointing out to Amazon which eBooks are riding the BotMobile (I'm totally stealing this word! - :D ) if they also follow certain other patterns.
Ooh, there's a thought. Easy enough to go down the BookBub mail each day to see what should be there at those quantities. Hmmm...
 

·
Registered
Joined
·
664 Posts
Hi Phoenix, here's one for you:

ASIN: B00M30BF5Y
1117 downloads on Weds on .com only
924 current rank as of 6:30AM
Permafree

Please keep us posted of your findings :)
 

·
Registered
Joined
·
983 Posts
6:30 AM Thursday morning, EDT.

ASIN B00GDDAT5E
1403 downloads Wednesday (.com only, this title only)
#715 Free in Kindle Store (See Top 100 Free in Kindle Store)
Select
 

·
Registered
Joined
·
2,168 Posts
I don't know the time conversions (I'm in Australia), but my Permafree is in the SPRT promo, and it's ranking right now is 814 on .com. I've had 267 downloads yesterday on .com, and 14 today.

It usually moves 10-20 copies a day, but there was a spike of 100 about 3 days ago. ASIN is B011LKBENE.
 

·
Registered
Joined
·
4,279 Posts
Discussion Starter · #9 ·
**********
 

·
Registered
Joined
·
2,168 Posts
Yeah, mine seems really out of sync with the other numbers that are being reported. No idea why.

It does occur to me to wonder if the free downloads are being reported properly?
 

·
Registered
Joined
·
4,279 Posts
Discussion Starter · #11 ·
**********
 

·
Registered
Joined
·
664 Posts
My book is quite old, released summer 2014. Could that be a factor? I did only make it permafree this month. It was in Select previously for about a year.
 

·
Registered
Joined
·
274 Posts
PhoenixS said:
Pfft. :-*

Correct on the procedure., Lindsay. Although BookBub'd books are probably not the best candidates for this. Massive enough downloads are breaking through whatever the issue is. So BB books are still hitting in the Top 10. What is impossible to know when there is that much competition for the top spots is what and where the true hit is.

For instance, a couple of weeks ago, we had a BB book on Day 2 of being Select free. On Day 1, the book had 1028 .com downloads. Best rank until the BookBub mail went out was #1594. Obviously this book was affected. It had 32,200 .com DLs on Day 2. It hit a rank of #4. Were the previous 1000 downloads still not being seen by the algo? If they HAD been seen, would the rank have been #3? Well, at that level, it's impossible to know because we don't know the circumstances of Book #2 or #3. Like with the Paid rank, once you're in to the Top 10, it's pretty much a free-for-all. #5 could have 5000 sales, while #4 could have 10,000. Between #20 and 100, the list is MUCH less volatile and much more predictable.

Do share your numbers, because all data is instructive! I just don't think it will be AS instructive in this case.
Gotcha! I'll keep an eye out when I run promos on the smaller sites.

Here's the BB gorilla:

- Your book's ASIN*

B01FWOW72W

- # of downloads on .com only

25,000

- Current rank on .com

2 at 8am PST this morning

- Select of Permafree

Select
 

·
Registered
Joined
·
4,279 Posts
Discussion Starter · #14 ·
LindsayBuroker said:
Gotcha! I'll keep an eye out when I run promos on the smaller sites.
...
Thanks, Lindsay! You probably noted that within an hour or so of your rank capture, no fewer than FIVE titles leapfrogged yours into #1-5 Free positions. FIVE titles with zilch promo anywhere for the exceptionally popular permafree (only, you know, not suddenly that exceptionally popular) and the 4 others that are Select that only went free 6 hours before they hit in the Top 5. They're falling in rank now (except the #1 title), although still in the Top 10. Which means there are 6 botted books (IMO) that are now in front of yours on the chart.
_______

Back in a few minutes with the first report-out of what yesterday's ranks:DLs looked like compared to what they would have looked like in early September...
 

·
Registered
Joined
·
4,279 Posts
Discussion Starter · #15 ·
Here's the first round of reports with 44 data points now. I've excluded the books that had BB runs. The chart on the left has the date of the reported downloads, the number of downloads that day, the actual rank captured after the reporting lag, and what that rank would have/should have been prior to Sept 12, which is now when I have the earliest report of wonkiness.

The top graph on the right (NEW PROFILE) shows the actual rank-to-download numbers. See how it's a very jagged, angry graph in dire need of anti-anxiety meds?

The bottom graph (OLD PROFILE) shows the nice, smooth curve that should be produced by a happy, well-adjusted rank calculation.

____

**********
 

·
Registered
Joined
·
3,171 Posts
Phoenix, are you wedded to the 15th September date as the starting point? Because I noticed this effect on a freebie on 10th September. Details, in case they're useful:

ASIN: B01H0TO1Q2
Downloads: 2593
Best free rank: #229
Book in Select, 1st day free.

Whereas a virtually identical promo in July, with 3244 d/l hit a rank of #36
 

·
Registered
Joined
·
4,279 Posts
Discussion Starter · #17 ·
PaulineMRoss said:
Phoenix, are you wedded to the 15th September date as the starting point? Because I noticed this effect on a freebie on 10th September. Details, in case they're useful:

ASIN: B01H0TO1Q2
Downloads: 2593
Best free rank: #229
Book in Select, 1st day free.

Whereas a virtually identical promo in July, with 3244 d/l hit a rank of #36
Not wedded to the 15th at all! It's simply a matter of finding the folk like you who had promos earlier (I had a free run that started on Sept 15 so that was my benchmark date) who kept track and know what a free run should do numbers-wise. In fact, I revised that to the 12th in the post right above yours. I'll now revise to the 10th. And yes, these results are useful and will be included. Thank you!
 

·
Registered
Joined
·
53 Posts
This explains something I was wondering about a lot. I don't have hard data to contribute this moment, but I will in a couple days, after Patty's Promo.

I just made my first in series permafree on the 20th, late at night. The next day, with no promo whatsoever, it shot to 167 downloads on .com. My rank never even broke 5,000. At the time I was ecstatic, but there were no follow-up sales, and I now believe that some bot used my book to camouflage their own activities. At any rate, I spiked my own downloads in a BKnights promo on the 26th, got 184 downloads on .com, only made it up to 4k or so this time. According to your chart, I should have hit 1000 at least, but I never even got close. I had no idea what it took to reach those, as I've never done free before.

Today I don't have a promo running, but have more downloads than yesterday. My rank has fallen 500 spots. Promo downloads will probably start tomorrow as a lot of emails go out early (including mine), so I will post my number then. Thanks for spotting this and letting people know. I hope I can help.
 

·
Registered
Joined
·
2,168 Posts
I thought the free charts were based purely on downloads, so the higher the number of downloads, the higher you we on the charts. More like the bestseller charts than the pop lists. If so, they're not run on an algorithm, as such. Or they weren't.

Leads me to wonder if they're excluding some kinds of downloads from the count. No idea what though.
 

·
Registered
Joined
·
26,676 Posts
Rinelle Grey said:
I thought the free charts were based purely on downloads, so the higher the number of downloads, the higher you we on the charts. More like the bestseller charts than the pop lists. If so, they're not run on an algorithm, as such. Or they weren't.

Leads me to wonder if they're excluding some kinds of downloads from the count. No idea what though.
Maybe it's another way to cut down on freeloads.
 
1 - 20 of 99 Posts
Top