State of retail publishing 2017

[ZodWallop]

Originally Posted by j.p.s

Your reference stated usually 10, you also used the plural, so even for a base of 2, that is at least a factor of 4, and your post implied Scalzi himself used the phrase "orders of magnitude".

Please let it go. Does anyone really care about the definition of 'orders of magnitude' at this point?

[pwalker8]

Originally Posted by sjfan

He does reference it, to point out what it misses. Most of the numbers Data Guy cites come from his Bookstat service, which was founded in large part to address the holes in Bookscan's tracking; most notably it certainly does at least attempt to cover ebook sales via its scraping methodology.

It's certainly legitimate to discuss how accurate his numbers are, but it's a very different system from the Bookscan one that Scalzi is criticizing.

Scalzi was specifically references the differences in the hardback sales verses the estimate. His comments about the various streams (paper, ebook and audiobook) are also important, but from a different set of blog posts.

Fine, if you want to say that Bookstat is more accurate than Bookscan, what were the differences in estimate between Bookstat and Bookscan for Scalzi's book? Scalzi references that screen scraping from Amazon isn't particularly accurate in other blog posts.

[shalym]

Originally Posted by pwalker8

Scalzi was specifically references the differences in the hardback sales verses the estimate. His comments about the various streams (paper, ebook and audiobook) are also important, but from a different set of blog posts.

Fine, if you want to say that Bookstat is more accurate than Bookscan, what were the differences in estimate between Bookstat and Bookscan for Scalzi's book? Scalzi references that screen scraping from Amazon isn't particularly accurate in other blog posts.

I'm confused...does Dataguy mention Scalzi's books explicitly? If not, how in the world does ONE data point have anything to do with a statistical analysis of the entire publishing industry?

Shari

[pwalker8]

Originally Posted by shalym

I'm confused...does Dataguy mention Scalzi's books explicitly? If not, how in the world does ONE data point have anything to do with a statistical analysis of the entire publishing industry?

Shari

Because it's one of the few real data points we actually have. Models have to match reality to be useful. If they don't match reality, well then what's the point?

What they are trying to do is provide a model of book sales. When the model provides results that prove false in one case, then the model isn't likely to be a reliable guide to actual book sales in other cases. In theory, you can claim that that result was an outier, but you can only do that if you have a lot of other actual data that prove to be correct. There are very few authors or publishers who will publish their actual sales figures. If I see other authors who say that these models accurately estimated their sales figures, I'll reconsider my position. So far, I haven't seen that.

[darryl]

@pwalker. So far you seem to be thoroughly confused. You can't seem to tell the difference between Bookscan, Bookstat and Author Earnings. You do seem to have construed Scalzi's comments on Bookscan as applying to Bookstat and Author Earnings. To use your own statement, with the words in bold, italic and underlined added by me, you say:

When the model (Bookscan) provides results that prove false in one case, then the model (Bookscan) isn't likely to be a reliable guide to actual book sales in other cases.

Yet your argument seems to in fact be:

When the model (Bookscan) provides results that prove false in one case, then the model (Bookstat) isn't likely to be a reliable guide to actual book sales in other cases.

Plainly your argument is absurd given that Bookstat did not exist at the time and Author Earnings to the best of my recollection did not provide a breakup by author.

[pwalker8]

Originally Posted by darryl

@pwalker. So far you seem to be thoroughly confused. You can't seem to tell the difference between Bookscan, Bookstat and Author Earnings. You do seem to have construed Scalzi's comments on Bookscan as applying to Bookstat and Author Earnings. To use your own statement, with the words in bold, italic and underlined added by me, you say:

When the model (Bookscan) provides results that prove false in one case, then the model (Bookscan) isn't likely to be a reliable guide to actual book sales in other cases.

Yet your argument seems to in fact be:

When the model (Bookscan) provides results that prove false in one case, then the model (Bookstat) isn't likely to be a reliable guide to actual book sales in other cases.

Plainly your argument is absurd given that Bookstat did not exist at the time and Author Earnings to the best of my recollection did not provide a breakup by author.

The blog post that I pointed to was from this year. Just because I point out that authors say that the estimates are not accurate doesn't make me confused, it just means that I want a little proof before I accept some guy's model as gospel.

If you want to post from some authors who say "yes, his estimates of my sales were spot on" then I'm willing to look. So far I haven't seen any such posts. It's not enough to simply say "Oh well, this model is the only information that we publicly have, so let's just act like it's accurate". I get that people want real data to talk about and I get that people like to speculate. Heck, I even get that people want to accept models that confirm their gut feelings or personal believes. Doesn't make it accurate though.

We are conditioned to accept these sort of models as accurate. Most people accept political polls as accurate, no matter how often they prove wrong when the votes come in. Emotionally, we think they most be accurate because a lot of people pay a lot of money for them, and a lot of people act like they are accurate. Same for TV ratings and Best Seller Lists. Fun to talk about, and a lot of money gets spent based on it, but is it accurate? I rather doubt it.

[pdurrant]

Originally Posted by pwalker8

If you want to post from some authors who say "yes, his estimates of my sales were spot on" then I'm willing to look.

Since Bookstat don't break out numbers by author, that's hardly likely to happen, is it?

[darryl]

@pwalker. In the first sentence of you last post you refer to a July 2017 post you linked to earlier as being "this year". You do know we are now in the second month of 2018, don't you? And your post goes downhill from there.

You pointed to a single author, John Scalzi, as having said Bookscan estimates were not accurate, not Bookstat estimates which were not even around at the time. Seems you are still confused. I will let the rest of your post speak for itself. One piece of advice. It is better to acknowledge your error and move on. It is not the end of the world. We all have our prejudices which can blind us and we all get things wrong from time to time. You are dealing with intelligent people who can read your previous posts and see what you actually wrote.

[pwalker8]

Originally Posted by darryl

@pwalker. In the first sentence of you last post you refer to a July 2017 post you linked to earlier as being "this year". You do know we are now in the second month of 2018, don't you? And your post goes downhill from there.

You pointed to a single author, John Scalzi, as having said Bookscan estimates were not accurate, not Bookstat estimates which were not even around at the time. Seems you are still confused. I will let the rest of your post speak for itself. One piece of advice. It is better to acknowledge your error and move on. It is not the end of the world. We all have our prejudices which can blind us and we all get things wrong from time to time. You are dealing with intelligent people who can read your previous posts and see what you actually wrote.

Ok, now you are just trying to play internet gotcha games. If your main point is that Jan 2018 is technically a different year than fall of 2017, you really don't have much of a point. I know how the game is played. Let me know when you find authors talking about how incredibly accurate these various estimates are and I'll take a look.

[pdurrant]

Originally Posted by pwalker8

Let me know when you find authors talking about how incredibly accurate these various estimates are and I'll take a look.

Since Bookstat don't break out numbers by author, that's hardly likely to happen, is it?

[pwalker8]

Originally Posted by pdurrant

Since Bookstat don't break out numbers by author, that's hardly likely to happen, is it?

Then how would one know if it's remotely accurate? As I said, there are a lot of models that people take as the gospel without any sort of validation.

[DNSB]

Originally Posted by ZodWallop

Please let it go. Does anyone really care about the definition of 'orders of magnitude' at this point?

Me! I care!

[DuckieTigger]

Originally Posted by pdurrant

Since Bookstat don't break out numbers by author, that's hardly likely to happen, is it?

Since when can you see ANY data from BookStat unless you have a subscription with them. I cannot even find information on pricing.

And once you have a subscription it should allow breakup by individual author or title.

Unless Bookstat is not the same as bookstat.com.

[pdurrant]

Originally Posted by DuckieTigger

Since when can you see ANY data from BookStat unless you have a subscription with them. I cannot even find information on pricing.

Bookstat is the back end to the Author Earnings report. That's where you can find data from bookstat.com

[fjtorres]

Originally Posted by pdurrant

Bookstat is the back end to the Author Earnings report. That's where you can find data from bookstat.com

Author Earnings offers a free public, global view of the data.

Datastat sells full, real time, detailed data to "select" customers.
Where "select" = Publishing company with over $10M in annual revenues.

It most definitely offers title by title, author by author data and the tradpub establishment paying for it has no doubt about its accuracy. Neither do the dozens (hundreds?) of authors, indie and tradpub, who are upset at being both exposed by that data and being locked out of seeing it.

All the evidence, from Shatzkin citing it as gospel to the money it is generating to the concern from authors says the data is real enough. and so is the abstracted global view the public sees.

Deniers notwithstanding.

https://www.deanwesleysmith.com/new-...rnings-report/