Author Topic: A New AMS Thread  (Read 39300 times)  

Online A.G.B

  • Status: Lewis Carroll
  • **
  • Posts: 214
  • Oxford
    • View Profile
Re: A New AMS Thread
« Reply #525 on: February 13, 2018, 07:29:25 AM »
Lily,

I don't understand. I thought the consensus was that it's hard to get AMS to spend your money, and when it does, that is a great thing. Why not up the budget? Also, I have the exact question as baldricko. If we run multiple ads per book, how are we supposed to measure conversion rate on any single ad? I don't get authors running multiple ads. Do they mean they are running multiple ads, one after the other, but with only one ad running at a time, or are they running multiple ads at the same time? Again, if it's the latter, how do you track the conversion rate for any given ad? Cassie, can you kindly chime in?

If you're not in select, tracking ad performance is much easier as you can see sales associated with each ad. It's the borrows that confuse things.

Offline Andres

  • Status: Madeleine L'Engle
  • **
  • Posts: 63
    • View Profile
Re: A New AMS Thread
« Reply #526 on: February 13, 2018, 07:43:59 AM »
Thanks, but I am and will be in Select.

Andres Pelenur | facebook | twitter

Offline Cassie Leigh

  • Status: Scheherazade
  • *****
  • Posts: 1586
    • View Profile
Re: A New AMS Thread
« Reply #527 on: February 13, 2018, 07:48:47 AM »
Lily,

I don't understand. I thought the consensus was that it's hard to get AMS to spend your money, and when it does, that is a great thing. Why not up the budget? Also, I have the exact question as baldricko. If we run multiple ads per book, how are we supposed to measure conversion rate on any single ad? I don't get authors running multiple ads. Do they mean they are running multiple ads, one after the other, but with only one ad running at a time, or are they running multiple ads at the same time? Again, if it's the latter, how do you track the conversion rate for any given ad? Cassie, can you kindly chime in?

From my perspective, sometimes AMS spending your money is a bad thing. I've had ads that spent $40-$50 in the first few days and never showed as having a sale credited to them. Or showed as having one sale for $4.99 credited to them. And I wasn't seeing KU page reads to make up for that difference. So if you have a new ad that isn't targeted well, it can be costly. I know some think any impression is great, so they may disagree. For me if impressions aren't leading to clicks AND then leading to sales, I don't consider that a successful ad.

Usually it is hard to get AMS to spend your budget, but not always.Newer ads are one example. If I start a new ad on my first-in-series romance title that ad almost always will max out a $5 budget on the first day. I actually did that this week and by the end of the day it was up to a $40 budget for the day. (I upped it by $5 each time it maxed out. Total spend for the day ended up being about $26.) That doesn't mean it was a good ad, though. I paused that ad for a few days because it wasn't looking good in terms of sales/borrows.

I have about five ads that are currently maxing out their $5/day budget. Since I price my books at $4.99, if I have one sale on one of those books and see that the ad has hit its budget, I up the budget. But sometimes I hit the budget for a day and there are no sales and no rank boost (for a KU title) and no paperback sales. In those cases, I'm spending on ads but not making anything back. I'll usually bump it once to $10/day, but I won't go past that until I can see some sign that those clicks are resulting in borrows or sales. But if I'm seeing enough sales for a day to show that I'm making my money back on that ad spend? I'll keep bumping that budget throughout the day. (Those ads seem to max out around $40/day in ad spend.)

In terms of running multiple ads...For users of Product Display ads, it seems to be common to run multiple ads. I'm not sure if it was in this thread or the other, but there's one user on here who starts two to three new ads each day on the same books. In that situation, I'd assume you monitor total ad spend versus ebook and paperback sales and estimated borrows for that day or ad period and don't worry about individual ads. (If that were me, I'd probably use the billing history to give me a summary of ad spend for the period rather than monitor on an ad-by-ad basis.)

I don't use that approach, but I'd assume that you only care about an individual ad if it gets out of control and spends too much, but otherwise you don't care about individual ads. I could be wrong and hopefully someone who uses that approach can chime in. Ultimately, for books that aren't in KU, the dashboard will eventually show gross sales amounts credited to that ad and you can see individual ad performance that way. (With all the caveats that whole series purchases aren't reflected in that number and neither are KU page reads.) If you're in KU, then you need to adjust what Amazon tells you for your estimated page reads that also came from the ad.


8 Pen Names. Genres: Non-fiction, Speculative Fiction, Romance.
Blog

Online A.G.B

  • Status: Lewis Carroll
  • **
  • Posts: 214
  • Oxford
    • View Profile
Re: A New AMS Thread
« Reply #528 on: February 13, 2018, 07:53:55 AM »
Thanks, but I am and will be in Select.

I was responding to your question about how people track the impact of multiple ads on the one title. My answer is they are probably not in select (like me).  :D

Offline Cassie Leigh

  • Status: Scheherazade
  • *****
  • Posts: 1586
    • View Profile
Re: A New AMS Thread
« Reply #529 on: February 13, 2018, 08:02:54 AM »
A couple of questions.

1. Has anyone noticed if there's a more general consensus among AMS users favoring either multiple ads per book or single ads?

2. I think I might follow Cassie's advice and limit my ads to one per book, but if I do that, which ads do I pause. I suspect that those ads that have been steady performers over a long period of time are worth running even though they have steadily run out of steam, to the point the ACOS looks decidedly unremarkable. They could be pulling in reads... I know there's only one way to find out, but this has been a bad month for sales and I've also read from several sources that once you pause an AMS ad and later restart it traction has been lost. Any thoughts?

1. I think there are two schools of thought on that one. I tend to run Sponsored Product ads and favor running just one ad per book. I did have a period where a long-running ad was slowing down on me so what I did was start a second ad on that same book and then alternate the two ads throughout the day. I'd run one until it seemed to be dying off. Pause it. Run the other until it seemed to be dying off. Pause it. Etc. My experience in the past was that running two Sponsored Product ads at the same time for the same book wasn't as effective as running one good ad on that same book. AMS change, so that could've changed, but I've stayed with one ad per book running at a time.

The issue of ads interfering with each other doesn't seem to exist on the Product Display side and it doesn't seem to exist between PD and SP ads. So there is a group of AMS users who run a large number of PD ads on the same book and do well with that approach.

2. If you have multiple ads running on a book and want to go down to one ad only...

First, not all ads die off over time. I have one that's been going strong since September and is getting better instead of worse. I don't think that's a normal outcome, though. And I do "tend" my ads by pausing non-performing keywords, changing bids, adding new keywords, etc.

Second, even if an ad does die off, it can be revived. One approach is pause non-performing keywords, change bids, and add new keywords. Another thing that seems to work is pausing an ad for a bit. A day, a few days, a week. Also, outside promotion of a book that gives that book momentum can then feed into an AMS ad picking back up. I'm pretty sure the romance ad I had do really well last summer (until I stupidly took the book out of KU) was an ad that I had paused for a while and restarted when I did a free run on that book. I know that ad had been around for months before that free run without doing much at all and then went wild for about four months.

So having said that...If you have a long-running ad that has generally been profitable for you, I prefer to try to revive that ad rather than start a new one. In the same way rank in the Amazon store seems to have a historical component to it, I think the ads do, too. But if none of your ads were all that great, then maybe start a new one instead. (For the same reason, no history.) And which ad to keep is a hard call. I have some lower-bid ads that are more profitable per sale, but because they're lower bid ads the sales are also less frequent. So for me and what I want to do with my books I prefer to run ads that have a higher bid and maybe make less profit per sale, but have a chance of high enough ranking to get me visibility and organic sales, too. Which of those options matters more to you is really your call. (And remember that if the books are in KU that ACoS number is not telling the full story, so be wary of using it to make your decisions.)


8 Pen Names. Genres: Non-fiction, Speculative Fiction, Romance.
Blog

Online LilyBLily

  • Status: Scheherazade
  • *****
  • Posts: 1911
    • View Profile
Re: A New AMS Thread
« Reply #530 on: February 13, 2018, 08:04:01 AM »
Interestingly, that ad that supposedly was running out of its daily budget never did. I think I made the right choice. If AMS had said, "We'll show your ad to your new keywords only if you pony up more budget," I might have agreed. But the email is never on point. It's robotic and general rather than specific.

My ads do not max out during the day, but by late evening p.m. eastern time, sometimes one will. The dilemma always is whether bumping up the budget will grab a few more sales from the western part of our country. I could reduce the budget again each morning to see if it maxes out at the lower figure, but who has that much time?

Honestly, this is just like gambling. I like gambling only if I don't lose.

Offline Andres

  • Status: Madeleine L'Engle
  • **
  • Posts: 63
    • View Profile
Re: A New AMS Thread
« Reply #531 on: February 13, 2018, 08:24:25 AM »
Thanks Cassie et al for the detailed responses. I guess my confusion was on the point of not relying on the AMS dashboard for any sales data at all, given the reporting delays (assuming you are not in KU). So if you are looking at your KDP dashboard report, then you would not be able to determine which ad led to sales (if running multiple ads). This would apply even when running one sponsored ad and multiple PD ads, right? I thought the entire point of having to make the effort to run Excel worksheets is because we cannot rely on the data within AMS and must rely on the reporting within KDP?

Cassie, I'm enjoying your ADS book and your Excel book.

Andres Pelenur | facebook | twitter

Online Anarchist

  • Status: Arthur C Clarke
  • *****
  • Posts: 2330
  • Methodological individualist
    • View Profile
Re: A New AMS Thread
« Reply #532 on: February 13, 2018, 08:47:43 AM »
I run multiple SP ads per book to test copy.

I watch the ACoS for each ad. Once I receive a certain number of clicks, I can determine the winner. I use that insight when creating new ads.

Incidentally, in my experience, it's possible to prevent old ads from dying. I have 2-year-old ads that are still driving volume. I don't want to get into tactics. I'm just mentioning it to dissuade folks from accepting dubious assertions as facts.

Test everything. Remember, nothing trumps your own data.

"Opportunity is missed by most people because it is dressed in overalls and looks like work." - Thomas Edison

"Strategy without tactics is the slowest route to victory. Tactics without strategy is the noise before defeat." - Sun Tzu

Offline Andres

  • Status: Madeleine L'Engle
  • **
  • Posts: 63
    • View Profile
Re: A New AMS Thread
« Reply #533 on: February 13, 2018, 10:44:32 AM »
Cassie,

Reading your book on Excel for Authors has left my head spinning! (the fault's on me, not on you). However, for total beginners, would you consider a) releasing some ready made Excel sheets to those who bought your book ( eg: inserting a link deep inside the book to download a template Excel sheet that is set up), and b) a video mini-course with Screencast to at least help absolute beginners learn how to take the data from AMS and understand the most important metrics so we can at least function in this business? Without the visuals, I find following the Excel instructions quite challenging.

Andres Pelenur | facebook | twitter

Offline Cassie Leigh

  • Status: Scheherazade
  • *****
  • Posts: 1586
    • View Profile
Re: A New AMS Thread
« Reply #534 on: February 13, 2018, 01:46:10 PM »
Test everything. Remember, nothing trumps your own data.

This.

I thought the entire point of having to make the effort to run Excel worksheets is because we cannot rely on the data within AMS and must rely on the reporting within KDP?

Cassie, I'm enjoying your ADS book and your Excel book.

Glad you're enjoying the books.

I don't use the Excel spreadsheets for daily ad maintenance. You can usually do that with a glance at the KDP dashboard and CreateSpace dashboard (if applicable) as well as your book's rank (if it's in KU). I know some do use Excel to track daily ad performance since AMS doesn't break that out on a daily basis. The issue with the AMS dashboard is that the estimated sales are almost always delayed by a few days which means you can't use it to decide whether an ad that runs out of budget today is performing well. So it's not that those sales never show up (at least I see some eventually), it's that they aren't there in time to help with day-to-day decisions.

Cassie,

Reading your book on Excel for Authors has left my head spinning! (the fault's on me, not on you). However, for total beginners, would you consider a) releasing some ready made Excel sheets to those who bought your book ( eg: inserting a link deep inside the book to download a template Excel sheet that is set up), and b) a video mini-course with Screencast to at least help absolute beginners learn how to take the data from AMS and understand the most important metrics so we can at least function in this business?

I'm actually working on turning a number of my non-fiction titles into video courses right now, because Excel and AMS are both easier to understand when you can see what's being talked about. Hopefully AMS and Excel for Self-Publishers will be out by the end of the month. Shoot me your email and I can send you a promo code and link to the course when it's done and the Excel examples that are screenshot in the book. As part of the course landing page there'll be a few videos offered for free, so if there's a section of either book you're really struggling with, let me know and I can see about making that section one of the free videos. Although no guarantees. I have yet to upload a course, so don't know exactly how it all works.


8 Pen Names. Genres: Non-fiction, Speculative Fiction, Romance.
Blog

Offline baldricko

  • Status: Arthur Conan Doyle
  • ****
  • Posts: 793
  • Sydney, Australia
    • View Profile
Re: A New AMS Thread
« Reply #535 on: February 13, 2018, 02:20:28 PM »
Excellent information, Cassie. Thanks!!

I am running multiple ads per title. The last 4 ads I put up were all - in my opinion - much better than those previous. I based them on what I had learned from blurb changes and best performing tweets. Twitter is not a bad way to test AMS Ad copy in my opinion, but I am still trialing that method. Those last 4 ads have not earned any book sales as yet. I am wondering where they are appearing on the carousel. Way way back, by the look of things.

I'll switch off all ads with less impressive impression increases and leave only one ad per title and see how that goes.

A new trilogy is on the way, and it's something of a satire.
Gerard O'Neill | Website | Facebook | Twitter

Offline Philip Gibson

  • Status: Arthur C Clarke
  • *****
  • Posts: 2462
  • Gender: Male
  • Laos (Moved from Yorkshire)
  • Philip and Cookie the Dog
    • View Profile
    • Hashtag Histories
Re: A New AMS Thread
« Reply #536 on: February 14, 2018, 07:20:48 AM »
I do take the risk of using AMS sales data to monitor performance of individual keywords.  I do this in the belief that while the data are not ENTIRELY reliable, they are SOMEWHAT reliable.  And after all, they are the only way to track performance of individual ads and individual keywords (I run multiple ads for the same books and many of my keywords are used in multiple ads.)

I think the Amazon algorithms value, compare and place related ads based on a grouping of factors.  This can explain why I find that identical keywords in multiple ads continue to generate their own impressions, clicks and sales and do not simply cancel each other out.

My theory is that the Amazon algos evaluate: "keyword + bid + its attached ad copy + the historical click/sales performance of that keyword bundle" and assign placement based on those 4 factors.

My methodology is simple and fairly strict.  I start all ads off at a baseline keyword bid of $0.16 which I have found over time gives an average performance.  After two weeks or so, I start to adjust keyword bids based on the following:

If a keyword has 20+ clicks with no sales reported, I reduce the bid to $0.06.
If a keyword has 1 sale reported to it, and fewer than 20 clicks for that sale, I leave the bid alone.
If a keyword has 2+ sales reported, I raise the bid to $0.26 to gain more clicks for that "good" keyword.
When a keyword has 3+ sales reported to it, I raise the bid to $0.31 to generate more clicks for that "very good" keyword.

This has worked quite well for the past 12 months.  All my 18 current ads remain profitable as I continue to monitor closely and adjust bids.
« Last Edit: February 14, 2018, 09:35:43 AM by Philip Gibson »

What if there had been social media during important historical events?
Philip Gibson | Hashtag Histories | Word By Word Graded Readers for Children

Online LilyBLily

  • Status: Scheherazade
  • *****
  • Posts: 1911
    • View Profile
Re: A New AMS Thread
« Reply #537 on: February 14, 2018, 08:43:26 AM »
I do take the risk of using AMS sales data to monitor performance of individual keywords.  I do this in the belief that while the data are not ENTIRELY reliable, they are SOMEWHAT reliable.  And after all, they are the only way to track performance of individual ads and individual keywords (I run multiple ads for the same books and many of my keywords are used in multiple ads.)

I think the Amazon algorithms value, compare and place keywords based on a grouping of factors.  This can explain why I find that identical keywords in multiple ads continue to generate their own impressions, clicks and sales and do not simply cancel each other out.

My theory is that the Amazon algos evaluate: "keyword + bid + its attached ad copy + the historical click/sales performance of that keyword bundle" and assign placement based on those 4 factors.

My methodology is simple and fairly strict.  I start all ads off at a baseline keyword bid of $0.16 which I have found over time gives an average performance.  After two weeks or so, I start to adjust keyword bids based on the following:

If a keyword has 20+ clicks with no sales reported, I reduce the bid to $0.06.
If a keyword has 1 sale reported to it, and fewer than 20 clicks for that sale, I leave the bid alone.
If a keyword has 2+ sales reported, I raise the bid to $0.26 to gain more clicks for that "good" keyword.
When a keyword has 3+ sales reported to it, I raise the bid to $0.31 to generate more clicks for that "very good" keyword.

This has worked quite well for the past 12 months.  All my 18 current ads remain profitable as I continue to monitor closely and adjust bids.

Just revised a very large batch of keywords following your system. Hoping my ad will regain its mojo. I'll report back.

Online Anarchist

  • Status: Arthur C Clarke
  • *****
  • Posts: 2330
  • Methodological individualist
    • View Profile
Re: A New AMS Thread
« Reply #538 on: February 14, 2018, 08:58:03 AM »
If a keyword has 20+ clicks with no sales reported, I reduce the bid to $0.06.
If a keyword has 1 sale reported to it, and fewer than 20 clicks for that sale, I leave the bid alone.
If a keyword has 2+ sales reported, I raise the bid to $0.26 to gain more clicks for that "good" keyword.
When a keyword has 3+ sales reported to it, I raise the bid to $0.31 to generate more clicks for that "very good" keyword.

It's a terrific idea to have such rules in place. It simplifies routine audits.
"Opportunity is missed by most people because it is dressed in overalls and looks like work." - Thomas Edison

"Strategy without tactics is the slowest route to victory. Tactics without strategy is the noise before defeat." - Sun Tzu

Offline Philip Gibson

  • Status: Arthur C Clarke
  • *****
  • Posts: 2462
  • Gender: Male
  • Laos (Moved from Yorkshire)
  • Philip and Cookie the Dog
    • View Profile
    • Hashtag Histories
Re: A New AMS Thread
« Reply #539 on: February 14, 2018, 09:11:27 AM »
Just revised a very large batch of keywords following your system. Hoping my ad will regain its mojo. I'll report back.

The trick for me is being able to assign sales to a particular ad and then to a particular keyword.

Fortunately, when I began publishing, I followed the convention of pricing all my eBooks and paperbacks to end with 99 cents.  This means that when I start a campaign and see an ad or keyword sales figure ending with .99, I can attribute one sale to that ad or keyword.  Then, ads/keywords ending with .98, .97, .96, etc. can have 2, 3 and 4 sales attributed to them respectively.

For example, look at these AMS sales attributed to some of my recent ads taken at random:

$22.98
$68.94
$84.93
$52.95

From the above, I can attribute the following number of (eBooks or paperbacks) sold:

2 sales
6 sales
7 sales
5 sales

Of course, this means relying on the rather unreliable AMS sales data.  However, I have found that while AMS often under-reports sales, I have not seen them over-report sales, so I think this works quite well.

Philip
« Last Edit: February 14, 2018, 09:14:49 AM by Philip Gibson »

What if there had been social media during important historical events?
Philip Gibson | Hashtag Histories | Word By Word Graded Readers for Children

Offline Andres

  • Status: Madeleine L'Engle
  • **
  • Posts: 63
    • View Profile
Re: A New AMS Thread
« Reply #540 on: February 14, 2018, 10:18:26 AM »
On a related note, to those who have been using AMS ads successfully, how does your ROI compare to FB ads and Bookbub CPC ads? To put it differently, are you still using FB ads?

Andres Pelenur | facebook | twitter

Offline Philip Gibson

  • Status: Arthur C Clarke
  • *****
  • Posts: 2462
  • Gender: Male
  • Laos (Moved from Yorkshire)
  • Philip and Cookie the Dog
    • View Profile
    • Hashtag Histories
Re: A New AMS Thread
« Reply #541 on: February 14, 2018, 10:43:28 AM »
On a related note, to those who have been using AMS ads successfully, how does your ROI compare to FB ads and Bookbub CPC ads? To put it differently, are you still using FB ads?

I've given up on FB ads.  I never got a positive ROI with them and find them difficult to set up and manage compared to AMS ads.

What if there had been social media during important historical events?
Philip Gibson | Hashtag Histories | Word By Word Graded Readers for Children

Offline Philip Gibson

  • Status: Arthur C Clarke
  • *****
  • Posts: 2462
  • Gender: Male
  • Laos (Moved from Yorkshire)
  • Philip and Cookie the Dog
    • View Profile
    • Hashtag Histories
Re: A New AMS Thread
« Reply #542 on: February 14, 2018, 10:46:01 AM »
I believe that all clicks should NOT be seen as equal, and that keywords (like the people they connect to) could therefore be seen to have their own unique personalities and produce different buying tendencies. That is, 100 clicks from one keyword do not have the same value as 100 clicks from a different keyword.

Do people here agree?

Or not?
« Last Edit: February 14, 2018, 10:47:49 AM by Philip Gibson »

What if there had been social media during important historical events?
Philip Gibson | Hashtag Histories | Word By Word Graded Readers for Children

Offline Andres

  • Status: Madeleine L'Engle
  • **
  • Posts: 63
    • View Profile
Re: A New AMS Thread
« Reply #543 on: February 14, 2018, 11:58:57 AM »
Philip,

I'm too new to answer your question but I also wanted to know how many keywords you are using (ballpark) on your best performing ad? Thanks for your very helpful feedback.

Andres Pelenur | facebook | twitter

Offline Philip Gibson

  • Status: Arthur C Clarke
  • *****
  • Posts: 2462
  • Gender: Male
  • Laos (Moved from Yorkshire)
  • Philip and Cookie the Dog
    • View Profile
    • Hashtag Histories
Re: A New AMS Thread
« Reply #544 on: February 14, 2018, 12:32:14 PM »
Philip,

I'm too new to answer your question but I also wanted to know how many keywords you are using (ballpark) on your best performing ad? Thanks for your very helpful feedback.

My best performing ad has 54 keywords that were the best keywords taken from previous ads which had hundreds of keywords.  The ad has been running for five months.  27 of the keywords have generated sales.  The best keyword has generated 26 sales.  Others have generated between 2 and 23 sales. 

Philip

What if there had been social media during important historical events?
Philip Gibson | Hashtag Histories | Word By Word Graded Readers for Children

Online A.G.B

  • Status: Lewis Carroll
  • **
  • Posts: 214
  • Oxford
    • View Profile
Re: A New AMS Thread
« Reply #545 on: February 14, 2018, 12:34:38 PM »
I've given up on FB ads.  I never got a positive ROI with them and find them difficult to set up and manage compared to AMS ads.

I'm finding both Facebook and Bookbub ads far more difficult to turn a profit on than AMS ads, which is annoying as I'd like to focus some of my advertising budget on the other stores.

Offline Cassie Leigh

  • Status: Scheherazade
  • *****
  • Posts: 1586
    • View Profile
Re: A New AMS Thread
« Reply #546 on: February 14, 2018, 12:38:27 PM »
I only run BB CPC and FB ads when I'm doing some other sort of promo (like a Bookbub feature ad) and want to supplement that. But one of my goals for this year is to crack some form of wide advertising option to drive sales on the other platforms.

Phillip, agreed about the keywords. I looked at one of my best-performing ads yesterday. It had a group of keywords that were generating a purchase every 3-5 clicks and another group that were generating a purchase every 10-15 clicks. Obviously, I prefer the keywords that generate a purchase every 3-5 clicks, all other things being equal.


8 Pen Names. Genres: Non-fiction, Speculative Fiction, Romance.
Blog

Offline Andres

  • Status: Madeleine L'Engle
  • **
  • Posts: 63
    • View Profile
Re: A New AMS Thread
« Reply #547 on: February 14, 2018, 01:54:51 PM »
Philip's data is so interesting because it goes against the prevailing wisdom that you need hundreds of keywords for the ad to perform. Test, test, test...

Andres Pelenur | facebook | twitter

Online Anarchist

  • Status: Arthur C Clarke
  • *****
  • Posts: 2330
  • Methodological individualist
    • View Profile
Re: A New AMS Thread
« Reply #548 on: February 14, 2018, 02:16:33 PM »
Philip's data is so interesting because it goes against the prevailing wisdom that you need hundreds of keywords for the ad to perform. Test, test, test...

You only need a single keyword for an ad to perform. The challenge is finding that keyword. The most effective way to find high-converting keywords is to test a ton of stuff and see what sticks.

I'm using the title of a celebrity's book for one of my keywords. This book has nothing to do with my book nor niche. Nada. Zilch. Zippo. But it converts like mad (over several hundred clicks).

Today, I audited two ads for another of my books. 20,300+ clicks, 11.6 million+ impressions, and low ACoS on both ads. You wouldn't believe some of the keywords that are driving conversions for those ads (and no, not crap like "book book book"). I didn't believe it myself.

The lesson is to test as many keywords as you can. Identify and leverage the winners. Prune the losers. Do it fast.


« Last Edit: February 14, 2018, 02:18:27 PM by Anarchist »
"Opportunity is missed by most people because it is dressed in overalls and looks like work." - Thomas Edison

"Strategy without tactics is the slowest route to victory. Tactics without strategy is the noise before defeat." - Sun Tzu

Offline Andres

  • Status: Madeleine L'Engle
  • **
  • Posts: 63
    • View Profile
Re: A New AMS Thread
« Reply #549 on: February 14, 2018, 02:44:47 PM »
You only need a single keyword for an ad to perform. The challenge is finding that keyword. The most effective way to find high-converting keywords is to test a ton of stuff and see what sticks.

I'm using the title of a celebrity's book for one of my keywords. This book has nothing to do with my book nor niche. Nada. Zilch. Zippo. But it converts like mad (over several hundred clicks).

Today, I audited two ads for another of my books. 20,300+ clicks, 11.6 million+ impressions, and low ACoS on both ads. You wouldn't believe some of the keywords that are driving conversions for those ads (and no, not crap like "book book book"). I didn't believe it myself.

The lesson is to test as many keywords as you can. Identify and leverage the winners. Prune the losers. Do it fast.

Very informative but actually somewhat depressing. If completely unrelated keywords can drive the most conversions, then the entire ecosystem is basically a crapshoot.

Andres Pelenur | facebook | twitter