right now i got 1,000 ringtone related keywords ready and i got 10 different ad variations ready to test out. my question is, should i put all 1,000 keywords and 10 ads in 1 ad group or should i divide them up into like 5 ad groups or more? also if i should seperate them, should i seperate them by the type of keyword? like put all the cell phone brand keywords together, all the free ringtone ones together, and all the carrier ones together etc??
ad group keyword question
- Thread starter blueyam1234
- Start date
- Status
the rule of thumb is 1-20 keywords per ad group (the lower and more targetted the better). Seperate them into related groups. Each group should have at least 2 ad variations (split test and pick the best performer, then write split test again and so on)
So you've got 1,000 keywords to test. Awesome. So your budget is about $18,750 per day? Good stuff, you're on the right track then...
Oh hang on, your budget ISN'T that large?
The thing about having too many keywords to start a campaign is, the campaign will NEVER get off the ground. If your budget is below the figure above, (which I'm going to go ahead and assume it is), you need to TEST less keywords at any given moment. Especially in an ultra competitive niche like ringtones. You should focus your entire daily budget on a handful, let them get clicks and then determine if they're a keeper or not. If they're not, (the majority won't be), delete them and add new ones. Keep this up until you hit your desired net income daily. Then you can pretty much set the cruise for awhile and move onto other projects. (Checking back in if things start sliding.)
Oh hang on, your budget ISN'T that large?
The thing about having too many keywords to start a campaign is, the campaign will NEVER get off the ground. If your budget is below the figure above, (which I'm going to go ahead and assume it is), you need to TEST less keywords at any given moment. Especially in an ultra competitive niche like ringtones. You should focus your entire daily budget on a handful, let them get clicks and then determine if they're a keeper or not. If they're not, (the majority won't be), delete them and add new ones. Keep this up until you hit your desired net income daily. Then you can pretty much set the cruise for awhile and move onto other projects. (Checking back in if things start sliding.)
ok i divided them all up into ad groups and gave each ad group there own targetted ads. i got 2 ads for each group to split test. so im going to take be jeepers advice and not run all ad groups at the same time. i got about 40 ad groups with 10-20 keywords in each. i got verizon ringtone, cingular ringtone, t mobile ringtone, groups, free ringtone group, ringtone group, motorola group, nokia group, wallpaper keyword group, etc..there allt argetted. so be jeepers are you saying that i shouldnt run all 40 groups at the same time? i should just run maybe 5 groups at a time to see how they do , then if one sucks ill pause it and activate a different one?
Yep.
In a very competitive area like this you have to utilize your daily budget to its fullest. That means testing maybe one adgroup at at time. If you're testing 20 keywords then your budget is going to focus on those ones only. If you're trying to stuff more in there, you need a bigger budget to make up for it.
This way you will find out quickly if you're on the right track. Oh and don't bid on "free ringtones" because you will waste a lot of money that way.
In a very competitive area like this you have to utilize your daily budget to its fullest. That means testing maybe one adgroup at at time. If you're testing 20 keywords then your budget is going to focus on those ones only. If you're trying to stuff more in there, you need a bigger budget to make up for it.
This way you will find out quickly if you're on the right track. Oh and don't bid on "free ringtones" because you will waste a lot of money that way.
To add a thought to what beejeebers said, there's often a side benefit of testing a very small sample of keywords...
In many niches, what works for one group of keywords will work for another group of keywords.
For example, if you're testing the following keywords:
eminem ringtones
eminem ringtone
eminem verizon ringtones
eminem verizon ringtone
eminem sprint ringtones
eminem sprint ringtone
... then you can likely expect similar results for the following keywords:
50 cent ringtones
50 cent ringtone
50 cent verizon ringtones
50 cent verizon ringtone
50 cent sprint ringtones
50 cent sprint ringtone
The results (impressions, clicks, conversions, etc.) won't be a mirror image, but your limited test of the "eminem" keywords can at least provide a launching pad for a "50 cent" adgroup.
This doesn't work within every niche (and sometimes, the similarity that you see in your numbers come from remarkably different keywords within the same niche), but it's a good process to follow.
So, testing small batches of keywords will conserve your limited funds and enable you to follow a more targeted PPC process.
In many niches, what works for one group of keywords will work for another group of keywords.
For example, if you're testing the following keywords:
eminem ringtones
eminem ringtone
eminem verizon ringtones
eminem verizon ringtone
eminem sprint ringtones
eminem sprint ringtone
... then you can likely expect similar results for the following keywords:
50 cent ringtones
50 cent ringtone
50 cent verizon ringtones
50 cent verizon ringtone
50 cent sprint ringtones
50 cent sprint ringtone
The results (impressions, clicks, conversions, etc.) won't be a mirror image, but your limited test of the "eminem" keywords can at least provide a launching pad for a "50 cent" adgroup.
This doesn't work within every niche (and sometimes, the similarity that you see in your numbers come from remarkably different keywords within the same niche), but it's a good process to follow.
So, testing small batches of keywords will conserve your limited funds and enable you to follow a more targeted PPC process.
ok well i focused on 1 ad group, and i let them run all day until my daily budget was reached. them i went and looked at my 50 free ringtone related keywords, and my 6 diff ad variations, and i picked the keywords that i received clicks on, and i picked the best performing ad. i then paused all my bad ads and keywords and only left the good ones active. so i had my 10 best keywords that had clicks, and my best ad running for a day. i set the budget at 30$ and i let it run. i ended up getting 40 clicks, and 3 of them converted for leads so i spend 30 dollars and made 45$...i guess thats pretty good but how can i get more, spend more?. Also when i started all 10 keywords were OK quality score but now 3 of them are POOR. the thing is, those poor keywords are in the top 5 of my 10 performance wise. also im averaging position 7 for free ringtone, and position 14 for free ringtones, and position 3 for tones. how do i get on the first page, top 1? do i spend more, or do i just let it run for a few days
Be careful using the word "free" most merchants don't like this and you could lose your commissions.
As for how to manage it. What's your target ROI? You need that before you do anything.
As for how to manage it. What's your target ROI? You need that before you do anything.
this is weird because yesterday my ctr was .60%, and i got 3 leads with 43 clicks, but today my ctr is 1.10% and i got 30 clicks so far but i have no leads. all my stats improved but my leads went down
Ok then, 200% means for every dollar you want to earn back $3, keeping $2 as profit.
Now to figure out what you can spend per keyword, you need to know your payout, (which I'm sure you do). For ringtones, let's guess at $12 per sale.
So here's the assumptions and the max spends to achieve them:
Your Investment: $1
Payout: $12
Target ROI: 200%
Target Conversion Rate (From Your Day One): 0.60%
Based on those Numbers the following results:
Maximum Spend per Keyword is $4.00
Maximum Bid per Keyword is $0.02
Which means at a payout of $12 per sale, you can spend $4.00 on every keyword at $0.02 max bid at an estimated conversion rate of 0.60% to get a an roi of 200%.
Let's make sure we're right by checking the ROI again:
% ROI = (Net Profit/Spend) x 100
% ROI = (($12 - $4)/$4) x 100
% ROI = ($8/$4) x 100
% ROI = 2 x 100
ROI = 200%
So the numbers DO work out. One thing to consider, you don't have enough data yet to make any decisions. Your first and second day numbers are not enough here to be statistically significant. You need around 1,000 clicks to make sure you're numbers are legit from a mathematical perspective, (more is better though).
So, why did you get 3 sales on day one and 0 on day two? It was a fluke.
This is the reason advanced ppc marketers are willing to lose money for the first couple days. They want statistically significant numbers, which is more valuable to them then losing a couple hundred dollars initially. Once you have some rock solid stats, it's just a matter of going into "robot test mode" and following all your rules for max spend, max bid and pausing the crap and testing the new stuff. (Hint, this is where a good bid management script would work beautifully, if you had one...)
Now to figure out what you can spend per keyword, you need to know your payout, (which I'm sure you do). For ringtones, let's guess at $12 per sale.
So here's the assumptions and the max spends to achieve them:
Your Investment: $1
Payout: $12
Target ROI: 200%
Target Conversion Rate (From Your Day One): 0.60%
Based on those Numbers the following results:
Maximum Spend per Keyword is $4.00
Maximum Bid per Keyword is $0.02
Which means at a payout of $12 per sale, you can spend $4.00 on every keyword at $0.02 max bid at an estimated conversion rate of 0.60% to get a an roi of 200%.
Let's make sure we're right by checking the ROI again:
% ROI = (Net Profit/Spend) x 100
% ROI = (($12 - $4)/$4) x 100
% ROI = ($8/$4) x 100
% ROI = 2 x 100
ROI = 200%
So the numbers DO work out. One thing to consider, you don't have enough data yet to make any decisions. Your first and second day numbers are not enough here to be statistically significant. You need around 1,000 clicks to make sure you're numbers are legit from a mathematical perspective, (more is better though).
So, why did you get 3 sales on day one and 0 on day two? It was a fluke.
This is the reason advanced ppc marketers are willing to lose money for the first couple days. They want statistically significant numbers, which is more valuable to them then losing a couple hundred dollars initially. Once you have some rock solid stats, it's just a matter of going into "robot test mode" and following all your rules for max spend, max bid and pausing the crap and testing the new stuff. (Hint, this is where a good bid management script would work beautifully, if you had one...)
thats how many i can get throughout the whole day, im only getting like 4k impressions too...i guess if i get on the top 3 positions then ill get a whole lot more clicks.
Well depending on how much money you have to throw at it just do what you're doing time 20 or 50. For example take a niche like payday loans which I play in. Go and throw in 5k keywords in new campaign in a new test account as we don't want to taint our real account. Then bid 3-4$/click which should get you top positions for most of the words, at least for a day or two. Also make sure you have tracking on the other side so you can see what words convert. Plan on losing a good amount of money for a while. Depending on the niche. I think Diorex says they lose $100k setting up a new niche which I imagine a large part of is gathering PPC data to setup a great campaign.
So by doing this basically what we're looking to do is
1. Find the words that get impressions
2. Find the words that get clicks
3. Find the words that convert right off.
So for each situation:
1. If a word get's a lot of impressions but no clicks then you probably need to tweak your ad copy. So in your junk account delete the word from the main grounps and make it in it's own new campaign something like "focused" and setup and adgroup for it. Make a laser focused ad for that and see if you can get it to get traffic.
2. Here we're getting clicks but they're not converting into sales. This means you need to work on your landing page for these words. Maybe look in to some dynamic keyword insertion or making some small tweaks to add the keywords or something related to the keyword. Now keep in mind if your other page is converting for everything else don't go changing it. Make a new page that's pretty much the same with small changes.
3. And of course this is the best case scenario where we just move these keywords over to our real account and let them start making us money.
I'm no PPC expert I'm just experimenting and learning as I go. But this seems to be working for me. The key is the faster you can get data you're tracking well the faster you'll be profitable. If you can afford to invest big up front then you're probably going to get there a lot faster. Hunting word 20 at a time it's going to take you a long time to get volume for non-major words.
As you're throwing grains of sand into a jar rather then pouring it in by the bucket.
So by doing this basically what we're looking to do is
1. Find the words that get impressions
2. Find the words that get clicks
3. Find the words that convert right off.
So for each situation:
1. If a word get's a lot of impressions but no clicks then you probably need to tweak your ad copy. So in your junk account delete the word from the main grounps and make it in it's own new campaign something like "focused" and setup and adgroup for it. Make a laser focused ad for that and see if you can get it to get traffic.
2. Here we're getting clicks but they're not converting into sales. This means you need to work on your landing page for these words. Maybe look in to some dynamic keyword insertion or making some small tweaks to add the keywords or something related to the keyword. Now keep in mind if your other page is converting for everything else don't go changing it. Make a new page that's pretty much the same with small changes.
3. And of course this is the best case scenario where we just move these keywords over to our real account and let them start making us money.
I'm no PPC expert I'm just experimenting and learning as I go. But this seems to be working for me. The key is the faster you can get data you're tracking well the faster you'll be profitable. If you can afford to invest big up front then you're probably going to get there a lot faster. Hunting word 20 at a time it's going to take you a long time to get volume for non-major words.
As you're throwing grains of sand into a jar rather then pouring it in by the bucket.
Excellent post Smaxor, I was targetting mine to someone who doesn't have a big budget to play with yet still wants to get into an ultra competitive niche. It will take them a lot longer to get up to speed, but it's possible.
- Status