AMS advice for the comically inept

[Simon Haynes]

I've been looking into the "Campaign bidding strategy" on a few of my AMS ads. I usually can't decide between 'Dynamic bids - down only' and 'Fixed Bids', preferring the former. (Sounds like a money-saver.)

At times I've run all my ads one or the other.

Anyway, when you read their help (I know, I know), it says that their bid adjustments are based on ACOS, of all things.  So, if you have lousy ACOS for a particular keyword/target, it will bid lower thinking it's not going to convert.

That's all well and good, but as we all know ACOS is about as accurate as a U-shaped arrow. On top of that, if I'm targeting hand-picked product ASINs, I don't want them cutting my bids by 80% because their algos think I'm not a good match.

So, I'm experimenting with fixed bids on closely targeted ads, and dynamic up/down on author/title/search keywords which could end up anywhere.

BTW the popularity for many of my titles has just climbed a lot on the scifi humour list in the USA. Whereas last week I had two books in the first 20 pages, I now have 15. (This list only seems to update once a week, and the latest was today.)

[LilyBLily]

Thanks for doing the investigating. As I understand it, I'd have to choose dynamic or fixed per ad, not per keyword, right? So if I have some suspected meh keywords they should start at fixed and eventually be eliminated?

[Lu Kudzoza]

I use fixed bids for the same reason you stated. I've done the research for books and authors that are a good match. I want a little higher bid on those keywords as an attempt to train amazon that my books are a good match to theirs.

[notthatamanda]

Keep a really close eye on the up down bid spending. When I first started AMS I did well with bid plus. When they did away with that in favor of up down I picked that thinking it was the closest to bid plus and I was quite wrong about that. Up down burned money about as fast as a book bub ad in "spend as fast as possible" mode.

I was offered by Amazon free webinars next week on advertising, so I am doing the Intermediate and Advanced ones they offered. If I pick up any good info, I'll post it here.

[LilyBLily]

I've registered for those seminars, too.

[Simon Haynes]

I tried to watch the basic intro one, but got a web page error. After that I gave up.

Can the seminars be watched any time after they've aired? I don't do live stuff, mostly because of my time zone but also because I stopped living to other peoples schedules around 2005.

I've never used the 'up and down' bidding method. To me that sounds like a fast way to empty my bank accounts.

[Simon Haynes]

Ohh, now I remember the issue. I got this massive 'blocked script' error covering my web browser window, and assumed it was trying to use Adobe Flash.

Now it's just sitting there, spinning gently.

[notthatamanda]

Oh sorry, I thought you said you were using up down.  I only registered for the intermediate and advanced webinars, but if they mention you can watch after I'll let you know.

[Simon Haynes]

Thanks for that!  I'm watching the intro one now. Slow going with lots of pauses and they don't have a 'speed up' button, alas.

I will experiment with Up and Down eventually. I should do it for one series and see whether it leads to better results.

[Hopscotch]

I'm having a bumpy ride using Bryan's method.  Applying his teaching, I got quick upticks in sales/reads.  Now rollercoastering down.  Anyone else the same?  Should I augment his training w/another goat sacrifice (my herd is already pretty thin)? 

[notthatamanda]

I had a slow start, then some sales, then not, now a sale again today. I've only spent 36 bucks and both my books are profitable. If they weren't I have a 100 clicks on one book so I would be dialing the bids down but I don't need to. I consider this the info gathering stage and I am prepared to wait until January to have enough good keywords to start generation 2. As long as I am profitable, I am sticking with the plan.

I think it's a crappy time to be trying to build momentum with the holidays, covid, a lot of turmoil due to the big vote. I'm looking at the daily search terms report and keeping a list of what terms lead to clicks and sales.

[LilyBLily]

Most of my trial category ads have done nothing much and cost nothing much, but sales and reads have increased. One ad has cost just over $20 so far; it sells a series of six titles and they have been moving. That's the sole category ad for that title. Out of 34 category ads, only two have spent significantly over a dollar. One spent over $7 very fast with no sales and I realized it wasn't drawing the right sort of crowd and paused it. That's the problem with some niche books; they may rightly belong to a certain category that has been invaded by titles that palpably do not belong.

On balance the category ads are not an exciting success but neither are they a horrid failure. Technically, having spent just less than $40 on them and earned significantly more, they are profitable. There's only one problem: I did a BookFunnel joint promo this month for a couple of books in that series and some of my sales and reads may have come from the promo instead of from the category ad.

I haven't examined the keyword ads to see if there is anything new to learn from their data.

[Simon Haynes]

I have a new tool now which imports the Advertised Products Reports from AMS US/UK/etc. Then I import this and display the ad spend against matching ASINs in my sales analyser software.

Basically, I can see royalty + kenp income in one column, and ad spend in another. There's a net difference displayed too.  Instead of saying 'I spent too much on ads' I can now see which books are costing more to advertise than they're worth.

What I haven't done yet is to take all the income for a series and deduct the ad spend from that, instead of just deducting it from book 1. That would be more useful.

[Hopscotch]

  I don't know where my reads come from.  How can I find out?  The AMS "KENP read" column only accounts for a tenth of the reads I've had this month.  I'm not running any other ads, don't have a newsletter and my website gets few visits.  Is there any way to figure it out so I can better target readers?

[notthatamanda]

Have you gotten to the top 100 in any subcategories or popular lists? Supposedly AMS only tracks for two weeks after someone clicks on an ad. Could people have made a note of you a few weeks ago and just gotten to it now?

[Simon Haynes]

The sales and KENP read figures on the AMS dash are woefully inaccurate. The only thing you can do is try to correlate clicks in the AMS dash with changes in the rankings for your books. That will tell you if there were borrows.

Start from the assumption that you have no organic sales to begin with (or take the average book ranking, sales and KENP for the month before you start advertising.)

[Hopscotch]

Thank you, Simon and notthatamanda - very helpful.  Now have a new puzzle:  AMS today suspended one of my Bryan ads b/c the ad text violates current policy.  Puzzling b/c this is an autotargeted Standard ad w/o any ad text.  Zipped off an email to contactus for guidance.  Anyone else had this happen?

[Simon Haynes]

Not for an ad with no ad text. A long time ago I saw something about covers and/or book titles/subtitles sometimes triggering it though.  (One rule I read was 'no covers where there's a gun pointed at the reader'.)

[notthatamanda]

I think you just have to wait to hear from them. They have a tendency to use one error message for a bunch of different things and sometimes things get flagged incorrectly. You can always start the ad again and see if it goes through.

[LilyBLily]

Probably a glitch, although it used to be said that Amazon's weekend crew was more censorious than the weekday crew.

[Simon Haynes]

My impressions/clicks/CPC have been climbing to the point where I'm winding back some of my wilder bids. I'm keen to see whether impressions plummet, or whether I've now given AMS a thirst for my ad dollars and they continue to show my ads despite the lower bids.

[idontknowyet]

I've been looking into the "Campaign bidding strategy" on a few of my AMS ads. I usually can't decide between 'Dynamic bids - down only' and 'Fixed Bids', preferring the former. (Sounds like a money-saver.)

At times I've run all my ads one or the other.

Anyway, when you read their help (I know, I know), it says that their bid adjustments are based on ACOS, of all things.  So, if you have lousy ACOS for a particular keyword/target, it will bid lower thinking it's not going to convert.

That's all well and good, but as we all know ACOS is about as accurate as a U-shaped arrow. On top of that, if I'm targeting hand-picked product ASINs, I don't want them cutting my bids by 80% because their algos think I'm not a good match.

So, I'm experimenting with fixed bids on closely targeted ads, and dynamic up/down on author/title/search keywords which could end up anywhere.

BTW the popularity for many of my titles has just climbed a lot on the scifi humour list in the USA. Whereas last week I had two books in the first 20 pages, I now have 15. (This list only seems to update once a week, and the latest was today.)

I read a post about using fixed bids to get amazon to see where you are engaging with the audience. By having a higher bid for a while you tend to get more clicks.

[Simon Haynes]

Certainly getting lots of clicks.  At this rate I'll finally be able to analyse the results, then pick out the targets which worked best.

Wouldn't it be nice if the Orders and KENP on the AMS dash actually worked?

[Simon Haynes]

Just found a popup on my AMS dash (US only) - introducing estimated KENP royalties.  I enabled the column, and it does what it says.

However, it's just the existing unreliable* KENPC x last month's KENP rate, so it's not that useful.


* I have noticed more page reads on the AMS dash recently, spread across a number of ads, but then I'm spending more so perhaps that's normal.

[LilyBLily]

I've re-upped my top-performing category ads for a short period and made them fixed bid only to see what happens with that variable. I've also initiated the fixed bid tactic for a few other ads. Another two-week experiment.

[Simon Haynes]

I'm going to create a few second gen ads soon, using the best keywords from the first round. (The ones that actually got impressions...)

[notthatamanda]

I haven't had a sale at all in five days so I'm all about chocolate and cheese.

[LilyBLily]

I haven't had a sale at all in five days so I'm all about chocolate and cheese.

 

[notthatamanda]

Yep Lily, you pretty much covered it.

Trying to remember in a normal year, mid Nov through Jan 1 would be pretty slow anyway.

Maybe I'll see you in webinar comments later.

[Simon Haynes]

After the bid adjustments, and having switched off a few underperforming keywords, I've managed to get my overall CTR (US store) from 0.19% over the previous 7 days up to 0.32% now.

CPC is 0.41c average across all my ads.

I spend most of my time tinkering with ads that have few impressions and no clicks. If I look at the last ten in my list for the day, and fiddle with those, I assume that tomorrow will be the next-worst-10.  Keep doing that, and I'll gradually improve or cancel everything ;-)

[Simon Haynes]

It really does seem that in order to run efficient, usable ads, you first have to run a bunch of ineffient expensive ads and throw out or fix the worst offenders. It could have something to do with training the algos, too.


(Aside from tinkering with the bottom 10 ads, I also watch the top 5 most costly ads. If the sales/page reads for the books in those ads aren't showing the expected gains, I adjust the bids down.)

[notthatamanda]

My questions and answers from the AMS webinar I just watched.

Q: if you use a key word of say "coffee machine" and bid on broad, phrase and exact aren't you bidding against yourself? Why is that recommended? Thanks.
A:
Hi Amanda, that is a great question! As long as your campaigns are coming from the same account then you will not bid against each other and having the same keyword in with all three match types will allow you to see how that keyword perfoms on a match type specific level and you can adjust the bid accordingly! You may also notice that a keyword performs well with one match type but not very well with a different match type and then you could remove the keyword with the underperforming match type while still keeping the keyword which converts well for you
Q:If you accidentally use the same keyword, say "coffee pot" for "broad" in two different campaigns do they also not bid against each other since they are from the same account?
A:
That is correct, you can have the same keyword in multiple campaigns and as long as they are coming from the same account they will not compete with each other.
Q:
So back to my first question, about campaign budget. There is a lot of chatter on line that if your campaign budget is bigger, Amazon will take your ads more seriously and you'll get more impressions. But you (general you) are saying it is relevance + bid and budget doesn't matter, correct?
A:
The size of your budget is not related to how seriously we take your campaigns We do want your campaigns to run thorughout the day, but increasing your budget isn't the only way to do this - take the time to optimize your campaigns by lowering bids, adding negative keywords, adjusting your bidding strategy, etc. to help your budget last longer throughout the day

[LilyBLily]

Thanks. I got a phone call (one that I wanted) just when the Q&A started so I missed them all.

On the whole I thought the presentation was cogent but did not detail a lot that was new to me. However, it confirmed much that has only been speculation.

It'll be interesting to see if the next one, at 4 PM my time, has any surprises. 

[notthatamanda]

Oh you wouldn't have seen those. I typed them in and they answered me directly. They are supposedly going to email a link with the full presentation, which I may have to do for the 4. The other laptop is fighting with zoom and the dance studio had to go remote for two weeks so I may have to give up my laptop for that.

[notthatamanda]

Simon, what criteria are using that you want to go 2nd generation? I only have 9 sales so far so I don't think I'm ready.

I'm going to try a concentrated ad, one author, all their titles, and do broad, phrase and exact to see what that yields.

[Simon Haynes]

For me it's impressions.  When I have an ad with 150 keywords and I run it over a week, I find that the few clicks are spread out so they're often in single digits for ASIN/Keyword target. Not really much to go on. Impressions, however, tell a story.

So, if I have 150 keywords and 100 of them haven't even had an impression, I survey a few of them. If they're for popular authors/books, I assume my bid is too low. If they're relative unknowns with very low sales rank, I assume there's little traffic to those pages (AND my bids might be too low.)  I don't need to advertise on those keywords.

First job is to remove all of the latter. Then I can copy the keywords with decent impressions into a new ad, and bid higher on those.

I also increase bids on keywords I think should work (in the original ad) to see whether it gets them moving.

[notthatamanda]

Thanks for sharing.

[Simon Haynes]

It's so frustrating when the tools Amazon give us aren't up to the job.  The AMS dash doesn't report income accurately, and the only way to find out if someone is borrowing my books is to repeatedly load the book list on the author page for each supported store, to see whether the rankings are climbing.

(A borrow doesn't show up anyware except as a jump in sales rank. It's only visible as pagereads when people finish the book they're reading now, and maybe 2 or 3 after that, before they finally start on mine.)

As an experiment, I've been running ads at a higher total daily budget than I want to. It seems to be the only way to gather the data I need (CPC/CTR/effective targets)

Some days I see a big spend on ads, with just enough sales to justify the spend, and a KENP which is about the normal average, if a touch higher.

It's only when I look closer at the breakdown that I see something pretty significant: Today, despite a modest overall page read count, a large percentage of those reads are for 'first in series' books. In other words, the ones I'm advertising.  A couple of them are 2-3 times higher than the page reads for book 2 in the same series, which means something is bringing in those new readers - and it has to be the ads.

The sales ranks for those books don't support these figures, which could mean these folk borrowed the books days ago, when I first boosted my ad spend, and are just getting to them now.

Anyway, if I'm reading the data correctly, the readthrough could eventually turn the ads into decent earners. If not, I'm throwing money down the well.

Usual case, with advertising.

[LilyBLily]

I agree about the frustration. I got one of those "Your budget has run out. Triple it and the world is your oyster" emails from Amazon last night and I did triple it. The ACOS on that ad now is 559.36%. 

That'll teach me. In one of the Amazon webinars yesterday they encouraged us to add to our budgets when we get such messages. It has never, ever worked for my books. Never ever. One wonders what they are smoking.

[Simon Haynes]

If I get those messages I leave it paused but lower my bids. Next day, I assume it won't hit the budget. (Unless it's getting a great CPC and I can see lots of orders/good sales rank for that title. Then I would raise the limit.)

Most of my campaigns are set to $10, with some at $15 or $25.  I've only hit one limit (a 15) in the past 6 months.

I have to remind myself that if I see $xx spent in one day, it's spread across 7 different series starters and my non-fic title.  If someone had one series they could divide my ad budget by 8 and they'd probably consider it pitiful.

What I could do is to stop all ads except those for ONE book, and then increase the bids and budget. The only reason I haven't is because I write in 8 different genres/markets, and so I can potentially reach 8 different audiences (and make 8x more money, hah.)

[notthatamanda]

Lily, was it an auto ad?

They said in the seminar the other day that they like the ad to run all day, so I agree with Simon, lower the bids.

[notthatamanda]

They just gave us access to the brand page (in the US anyway). It does a landing page for you and you choose what to put on it. For example, I could do a page for each trilogy or all my romance (I think it said 8 products max). It showed you the landing page but I don't know what the customer is going to see in the sponsored ad. You pick the keywords.

[Simon Haynes]

Do you mean the brand page is available now through AMS, or they made it available to people who watched the vid?

[notthatamanda]

Hmm, well I watched the video, so I can't tell if everyone got it, or just people who watched the video.

[Hopscotch]

Pardon me if I do this first:   B/c, after applying Bryan's AMS magic and all your good advice for a month, I've achieved a consistent daily level of sales and reads.  BUT my AMS chart reports that only about 1/3 of those sales and about 1/5 of the reads are attributable to AMS.  Can anyone help me understand what's happening?

[Simon Haynes]

AMS doesn't report all the sales and reads generated by your ads. (Oh yeah, and as of a few weeks ago it doesn't report sales of your OTHER books generated by a click on your ad. So if someone clicks the ad then buys your entire series in one go, you'll only see a sale of the first book or - more likely - none at all.)

That's what makes it so frustrating.

By the way, they also say 'allow 2 weeks for data to show up'. Now, I've never figured out whether they mean the old data shows up in today's/yesterday's dash OR it changes the reports from 2 weeks ago.  I ought to take a screenshot of the AMS dash with the date set to 9 days ago and then check it again in 2 weeks.

[notthatamanda]

Different reasons. For example, if you run an ad and someone buys your book because of that ad, now your book is in also boughts with all of their also boughts. Now someone buys a book from the also boughts that will never be attributed to the ad. And that reader may never have found you with out the ad. I like Bryan's strategy of focusing on the overall profitability. When I did a lot of AMS in 2019 I had lots of sales that were never attributed to ads which was fine by me.

[Simon Haynes]

They just gave us access to the brand page (in the US anyway). It does a landing page for you and you choose what to put on it. For example, I could do a page for each trilogy or all my romance (I think it said 8 products max). It showed you the landing page but I don't know what the customer is going to see in the sponsored ad. You pick the keywords.

Do you mean this: https://advertising.amazon.com/resources/ad-specs/landing-pages


Ahh, never mind. When I clicked Create Campaign it now shows the brand option.

So I guess that means I need to create new ads instead of just retargeting the existing ones. Ouch.

[LilyBLily]

Just got another "You're about out of budget" email. It's a category ad with a $5 budget, and I was experimenting with fixed bids. I've switched to dynamic down only, since the ACOS is now 660.54%.

I think this is proof that fixed bids can run one out of money fast without producing great results. Or possibly the ads dashboard reports money spent faster than it does page reads or sales. In either case, it makes deciding whether to keep an ad going very dicey. This is not the only ad I'm running for this series, which complicates the situation further.

I have let most of my dozens of category ads expire. Interestingly, although the ads dashboard does not show any clicks for another series also in KU for which I have a category ad running, someone has been reading it. This may be more proof that the dashboard simply is not reporting information in a usefully prompt manner. 

This stuff hurts my brain.

[Simon Haynes]

Yes, you definitely have to ignore the orders, sales and pagereads on the AMS dash. Just pretend those columns don't exist.

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.  If you advertise more and those figures rise, especially for the first in series, it's working.