AMS advice for the comically inept

[notthatamanda]

Are you taking the titles and authors from the category ad search term report and creating keyword ads with those?

Category ads sold the most books for me, but they were expensive, with higher ACOS. Taking the books that got the most sales and clicks from those ads gave me great keywords for cheaper ads.

Also, what category? Rhetorical. You can try to narrow down the categories so you aren't using just a broad one.
Bid lower on Romance, higher on Historical Romance.

[Simon Haynes]

I'm testing a slightly different strategy for one of my ads. I've halved all the bids and changed the bid type to 'up and down' (the one which goes up to +100%) and I've also added a 10% loading to 'top of search' and 'product page'.

[Simon Haynes]

I'm testing a slightly different strategy for one of my ads. I've halved all the bids and changed the bid type to 'up and down' (the one which goes up to +100%) and I've also added a 10% loading to 'top of search' and 'product page'.

Not the first time I've tried this - last time was months ago, but the ads never fired up. (Low bid was TOO low I think.)

Unfortunately this strategy relies on Amazon recognising which product pages to push the book onto.

[notthatamanda]

You got more balls than me. I tried up down when they first introduced it and lost my shirt, pants and underwear. Good luck.

[Simon Haynes]

Haha. Lucky I have my faithful dressing gown.

But if you halve the bids and allow Amazon to double them as required, isn't that a better outcome than setting the bids high in the first place and locking them to that amount?

(Don't answer - I'll soon find out.  I've been running the ad pre-changes for a week, so I'll have a before and after as a comparison.)

[Hopscotch]

If you're not advertising anywhere else, then you just have to assume all sales/borrows/pagereads/sales rank increases are due to the ads.

Okay, I think I've got it now:  Ignore the AMS dashboard - it's a poor tool to measure/calculate AMS expense vs profit b/c its reporting is neither timely nor complete.  Use the KDP dash instead - b/c it's more accurate/timely re: sales/reads - to calculate overall AMS success.  And to calculate down to book/series level.  However, in seeking out keywords that draw buyers, use AMS dash for impressions/clicks - understanding the imprecision of the AMS dash means reading those #s only as trendlines.  Or

[Simon Haynes]

I'd say that's right. Impressions and Clicks (and spend!) are the important parts of the dash.

If I have an ASIN-targeted ad where I'm paying for a lot of clicks, that's great because I've picked those ASINs as close matches for my books. If they're not buying/borrowing despite all the clicks, then I'll tweak the blurb/subtitle (or reduce the bids/remove that target)

If it's a keyword ad and it's getting lots of clicks, I'll check the reports very closely to make sure all the clicks aren't people searching for burglar alarms or books on DIY pet food or something.

[LilyBLily]

I keep adding to my list of negative targeting words and phrases. I am not paying for a click from someone looking for a "nasty western romance."

[notthatamanda]

If you're not advertising anywhere else, then you just have to assume all sales/borrows/pagereads/sales rank increases are due to the ads.

Okay, I think I've got it now:  Ignore the AMS dashboard - it's a poor tool to measure/calculate AMS expense vs profit b/c its reporting is neither timely nor complete.  Use the KDP dash instead - b/c it's more accurate/timely re: sales/reads - to calculate overall AMS success.  And to calculate down to book/series level.  However, in seeking out keywords that draw buyers, use AMS dash for impressions/clicks - understanding the imprecision of the AMS dash means reading those #s only as trendlines.  Or

And look at your search term report to see where the automatic ads are placing clicks. They are what the algos think are the most relevant and since relevance is key in getting impressions, and below bid impressions, you want to keep an eye on those. I run it daily.

[notthatamanda]

My questions and answers from the webinar today. I typed these into the chat. My screen froze up when the presentation was over and I didn't hear the live Q & A.

Is there an optimum number or range of keywords to use in a manual campaign? Does it matter?
A:
Hi Amanda, when you get started we recommend aiming for around 30 keywords but this can be more if you are a KDP advertiser

Thanks. Is too many keywords bad? I can go over a hundred in a campaign usually.
A:
as long as you keep checking to make sure that all keywords are driving sales. if you add to many you may see that your budget runs out pretty quickly.

[Simon Haynes]

Okay, they're certainly dispelling a few myths then.

I was holding off starting 2nd gen ads until I had lots of keywords, but maybe it's possible to fire them up with 30 after all.


PS I've been working on SalesScanner, and it will now show a chart of the last 30 days of royalties and pagereads, with averages, and also (new) it will draw another line on the chart indicating daily ad spend. (You can tell I spent 20 years writing share market charting software...)


The ad spend line requires downloading a specific report from AMS (or having it emailed daily, which is what I do) and then converting it using yet another program I wrote. I aim to roll that program into sales scanner.  At the moment it supports reports/ad spend from the US/UK/AU and CA markets.

Reason this is great is because I can see at a glance whether my royalties are climbing faster than the ad spend. (Recently a curbed a few bids, and while the royalties have kept increasing, the total ad spend dropped.)

When I say 'royalties' I'm including KENP in that. Since I've been heavily advertising all my 'book ones' I would expect KENP to have a long tail due to readthrough.  That's the one variable I don't have yet - whether that long tail will show my recent higher ad spend to be a mistake or not.

[Simon Haynes]

I've just updated the program so it'll display charts per market. This is a sample for one of them:

As you can see, ad spend jumped a lot yesterday, even with my lower bids.  I can also see I was more profitable a month ago with minimal ads, but the recent ad spend has definitely reversed a slump.

(It's not colour-blindness friendly at the moment, but the bars are green when they're higher than the previous day)

[Simon Haynes]

Here's the KENP chart for the same period and market.  It's what you might call a V-shaped recovery.

And as I said, the ad spend going up is fine as long as these page reads (mostly for first in series) translate to readthrough. Behind the 7 books I'm advertising are 23 others, all lined up and ready to read.

[notthatamanda]

I'm going to be trying some general word ads soon versus authors and titles. I had some luck with those in the past, but the bids are usually pretty high on those and I'm sticking with the 30 cent maximum for the time being.

[notthatamanda]

They just changed the copy ad completely. I wanted to copy an ad I did and use it for a different book but instead of putting you into the new campaign page it just launches it (you can choose to pause it) and you can't change the book. I can export the keywords but man, I don't see the point of this. The old way was better. Maybe has something to do with that bulk stuff.

[Simon Haynes]

If the original ad had a blurb, then I'm guessing that's why you can't change the advertised book? Could you ever? (Can't recall)

With a blurb-free ad you can just add another book and untick the first.

[notthatamanda]

Nope, this ad didn't have a blurb. But I'm pretty sure I just used to copy the ad and it would shoot you into the new ad page and then you could change everything. I exported the keywords and negative keywords and copied them back in. Took 10 minutes (my laptop is slow). Old way would have taken a minute but it's done. Will all be worth it when I get 4 clicks and no sales over the next month.

[Simon Haynes]

Now you're just being flat out over-optimistic.

[Simon Haynes]

Still, it's probably a better outcome than paying twice as much for bids and ending up with 'negative royalties'

[notthatamanda]

Well if I'm the optimistic one, we're in trouble.

Thinking it over, I am pretty PO-ed about it. The two main reasons I would copy an ad would be to change the creative or to change the book. Now I can't do either. I'm not sure what the point is of this change is. Everything you can change when you copy an ad now you could always change in the original ad anyway. And since I learned that I can't compete against myself, I was prepared to test a lot more copy head to head on the same keywords, which will now be more of a PITA to set up. 

Edit - here are some interesting results. One author and all her titles for keywords. Identical ads with 6 different hooks and one with no hook.  Impressions so far:

Hook 5 - 48
Hook 6 - 84
No Hook - 57

No clicks yet, but the first hurdle is getting the impressions and the algos clearly have preferences. I have another author to set up an ad for but I will still run all 6 hooks and none, to see if there is any difference between what the algos prefer for a different author, or if it is pretty standard.

[Hopscotch]

Am I wrong to think that, very generally, 1,000 impressions should generate 10 clicks that produce 1 sale?

[Simon Haynes]

Well if I'm the optimistic one, we're in trouble.

Thinking it over, I am pretty PO-ed about it. The two main reasons I would copy an ad would be to change the creative or to change the book. Now I can't do either. I'm not sure what the point is of this change is. Everything you can change when you copy an ad now you could always change in the original ad anyway. And since I learned that I can't compete against myself, I was prepared to test a lot more copy head to head on the same keywords, which will now be more of a PITA to set up. 

Edit - here are some interesting results. One author and all her titles for keywords. Identical ads with 6 different hooks and one with no hook.  Impressions so far:

Hook 5 - 48
Hook 6 - 84
No Hook - 57

No clicks yet, but the first hurdle is getting the impressions and the algos clearly have preferences. I have another author to set up an ad for but I will still run all 6 hooks and none, to see if there is any difference between what the algos prefer for a different author, or if it is pretty standard.

Common wisdom is that copying ads is not the way to go. (For some reason the copy carries baggage from the original - e.g. performance figures.)  I know David Gaughran says not to, and Bryan Cohen does too.  I don't copy ads for that reason.

[Simon Haynes]

Am I wrong to think that, very generally, 1,000 impressions should generate 10 clicks that produce 1 sale?

Hard to say, as it's impossible to directly measure sales.

Picking one of my books in one marketplace from yesterday, the set of 10 ads generated 9000 impressions, 19 clicks, 2 orders and 128 page reads. My dash shows 1 sale and 300 page reads for that book in that marketplace - but I guess it's possible someone refunded a copy. (Nope, just checked and I've not had a any returns for that one all month.)

So, that second sale probably took place some time over the past 14 days and has only just been credited to the ad. Annoying.


A different book yesterday had 800 impressions, 1 click, no sales showing but there's one sale for that one in my figs.

[Simon Haynes]

The other day I set up two of the new brand ads. I got an error message for one, but when I clicked submit a second time it went through.

Anyway, no impressions yet, and I've been raising my bids a little at a time to try and get them to fire up.

Yesterday I created a third Brand ad for a different series. It was rejected an hour later with a message saying i must provide an author profile pic.

I opened up the draft (they helpfully converted my ad to a draft so I could edit it), and sure enough there's a pic there. I mean, it stayed there from the original two ads and they were both accepted.

So, clicked submit.  Error: You've already used that campaign name. I added a 2 to the campaign name and resubmitted.

An hour later: rejected. No profile pic.

This time I deleted the profile pic, uploaded it again, cropped it, changed the ad title and submitted again.

Ding! No profile pic. Third rejection. (Each time I went to the feedback page linked from the ad and clicked the 'I don't understand' option.)

By now I'm getting a bit fed up, so I left everything alone and changed the ad title:

ITS GOT A PROFILE PIC FGS.

and clicked submit. I regretted it immediately, because I don't usually do that sort of thing. I consoled myself with the thought that (hopefully) I'm only railing against a dim-witted AI.

Anyway, this morning the ad was accepted, so I quickly changed the title and went on with my life.

[R. C.]

Adding fuel to fire...

The fears are being realized.

The Sponsored Brands AD was approved but has ONE impression in just under 24 hours.

I'll bump the bid and see what happens.

Cheers,
R.C.

[Simon Haynes]

Yep, it's a winner takes all situation.

[notthatamanda]

You guys think they are favoring the sponsored brand ads now?

[Simon Haynes]

Just a thought, but has anyone seen a brand ad with the three books in yet? (According to the setup/help they don't show on product pages but in search results. Therefore, my targeting of ASINS is probably a waste of time.)

[notthatamanda]

I'm not even sure where to look. Occasionally I a sponsored ad on a book page under the book description, all by itself, not in a carousel. Maybe it would be there, but those are rare.

[Simon Haynes]

It says they show in search results, and doesn't mention product pages.

I added some categories along with my ASIN targets, because we all know how well the former work...

[notthatamanda]

Well you can tell me what you used (PM me if you want) and I'll search and see if anything comes up, yours or anyone else's. I don't think our categories are terribly similar but I promise not to use your keywords.

[LilyBLily]

Meanwhile, my flirtation with fixed bids has ended. Amazon spent my budgets extremely quickly, and alas, sales and reads have not equaled the ad costs. Back they all go to dynamic down only.

[notthatamanda]

Ick, sorry. Better luck with dynamic down.

[LilyBLily]

I'm not risking a lot of cash, but I don't like seeing an ACOS of 672.07%. Hard to believe the numbers can go that high! 

[Simon Haynes]

Yep, I'm finding the fixed bids too rich, especially later in the day when AMS seems to delight in spending as much as possible.

I am trying the up/down bids though. I set the keyword bids to half what I normally would - that way, worst case is I'm paying my normal max. I also added a healthy percentage for top of search and product page placements.

[notthatamanda]

Now the freaking export isn't working. Typing in all the keywords manually again here I come. 

[Simon Haynes]

An up to date 30-day KENP chart. The last green bar isn't done yet - about 4 hours to go I think.

There's a visible upwards trend since I started advertising heavily and over-bidding. (That part isn't a surprise, except to prove that ads will generate pagereads, even if they eat up all your profit into the bargain)

What IS a surprise to me is that almost half that last green bar consists of page reads for my 'first in series' books - just 6 titlles out of 25+.  So, like I said yesterday, there's a chance those ads will continue to pay for themselves over weeks to come.

In the past I've never spent this much on AMS, and I've always pulled the pin after 2 or 3 days. This time I'm determined to soldier on and find out whether it's possible to get things to snowball.

I'm not completely nuts, though. I've been converting fixed bid ads to up/down ads and trimmed most of the bids in half in the process. As a result my ad spend is half what it's been recently. But then, of course, there are also half as many impressions and clicks ;-)

[Simon Haynes]

Well you can tell me what you used (PM me if you want) and I'll search and see if anything comes up, yours or anyone else's. I don't think our categories are terribly similar but I promise not to use your keywords.

Do you mean for the brand ads? I'm not using keywords there, just ASINs and (now) categories.  I did enter 'humorous fantasy' and saw an urban fantasy brand ad at the top, but further down on the first page were one or two of my book covers so I don't care (organic results, not paid.)  I repeated it for 'humorous science fiction' and saw a brand ad for an author with similar stuff to me, but I've seen his ads for months so I suspect he had an Advantage account. Again, 2 or 3 of my books showed up organically on the first page of results, so my keywords etc are working.

I think keyword ads are the way to go for these brand ads though.

[notthatamanda]

Yes I did, but it sounds like you got it covered. I still haven't run across any brand ads poking around on Amazon.

Update - so I started 5 ads at basically the same time, with the same keywords, bidding on broad, phrase and exact for each keyword. 4 different hooks and one ad with no creative. Default bid is 30 cents. Got my first click for 6 cents. So I think I bid against myself 15 times and it didn't affect what I paid.

[Simon Haynes]

Yes I did, but it sounds like you got it covered. I still haven't run across any brand ads poking around on Amazon.

If you go to amazon.com, select Kindle Store from the dropdown to the left of the search bar and enter Humorous Fantasy, what appears at the top of the search results? I get a brand ad (3 covers, a logo/photo to the left, etc)

[notthatamanda]

Humorous fantasy gets me a brand ad for a very dark looking epic saga series.

I tried humorous fiction and I got a brand ad for mafia romance.

Humorous science fiction gets me a brand ad for humorous science fiction.

So the ads are showing but the targeting seems a little off. I don't know if just not enough people are bidding on those spots yet cause the first two seem way off on relevance for me. Maybe it is a good time to try an brand ad for each of my trilogies.

Edit - did one for each of my trilogies. I'm sure romance will be very competitive. I bid 15 cents, don't expect a lot of placement.

[Simon Haynes]

That dark looking saga is meant to be a comedy, if it's the 'never shoot an ogre' one.  I just added all his ASINs to my targeting because I've never heard of it.

I'm doing a new pair of Brand ads now for keywords instead of categories.  I also deactivated all the ASINs from my manual targeting brand ads, because they never did anything no matter how much I bid.  I just left one category in each.

Once the new keyword brand ads are up I'm going raise the bid, refresh, raise the bid and refresh until they show.

[Simon Haynes]

Edit - did one for each of my trilogies. I'm sure romance will be very competitive. I bid 15 cents, don't expect a lot of placement.

I had to bid 10x that before I got any impressions at all (3).  That's with categories.

[notthatamanda]

I did categories for keywords, eg romance, contemporary romance. No impressions yet. I got to keep my costs down.

The one I'm getting in humorous fantasy is about an evil prince and crows. Book one has a best seller tag for "Teen & Young Adult Depression Fiction Books". Did not know that was a category, but it really doesn't sound like humor. Are you bidding on humorous fantasy? Cause your books would be a better match.

I have a vague idea about using AMS to test copy and translate that into more relevant blurbs but that would be a slog.

[Simon Haynes]

I'm still trying to get them to accept my keywords ads after about 20 attempts. Every time I submit, they reject them for not having an author profile image. (Which they have)

Yesterday I submitted the same ads over and over until they finally accepted them. Today I've resorted to emailing support.

So, at the moment I only have a category ad running.

Once the keywords ads go up I'll experiment with bids to see whether I can get my brand ad to appear. (I can bid a lot higher on humorous scifi because there are 10 novels in that series vs only 3 in the humorous fantasy series.)

[notthatamanda]

I had a tough time with one submission for the brand ad, but it took eventually. Also the review was longer. My sponsored products ads usually run immediately. Brand ads took overnight in review. I suspect it will take readers a while to look at them, or I might just be weird. I was seeing some but I would just scroll past them to get to what I expected to see. It took you pointing it out to me for me to actually look at them.

Edit - over 100 impressions for one of the brand ads and 39 for the other, bidding 15 cents. No clicks yet. I got a download of the first in series free on the first one, not credited to the ad. I really hope the landing page will make it clear to people that only the first book is free.

[notthatamanda]

Up to 247 on one of my brand ads. Man, I must have hit the sweet spot on relevancy.

If the book is free does anyone know if AMS counts that as a sale?

[Simon Haynes]

It will show up as an order, yes.  Congrats on the impressions!

I've seen a few impressions since I switched to categories instead of ASINs, and I resubmitted my keywords brand ads.

My ad spend cratered overnight - combination of reining in the biggest-spending keywords and tinkering around with others I suspect. (I changed a lot of my ads to up/down and much lower bids, but used bid+ to try and get the first page of the carousel plus product display ads.)

[Simon Haynes]

Okay, mine are getting clicks now.

If you edit the campaign and view the creative, you can click the 'Shop AUTHORNAME books' link to see your landing page.

[Simon Haynes]

You can see I'm managing to curb the ad spend (red line) but the royalties + pageread income are still climbing. (Bear in mind I have the 99c promo running)

The green bar on the right still has 3 hours of US sales to go. The tiny red bar to the right of that is the start of today's UK and AU sales. (Curse those timezones!)

The lower the red line and the taller the bars, the more profit.

My ad spend for today just leapt up (not displayed yet - figures come in tomorrow), so there might have been a delay in live reporting on AMS.