Andy Black
Help people. Get paid. Help more people.
Staff member
FASTLANE INSIDER
EPIC CONTRIBUTOR
Read Fastlane!
Speedway Pass
(Originally posted on the inside here.)
How To Get Profitable Or Fail Fast With AdWords
Here's how I often determine whether a project, niche, seed, or keyword has legs, or needs a bullet.
Note1: I use blanket bid prices throughout the steps below, meaning the bid price within a campaign for each ad group or keyword is identical. I can then judge how much supply/competition there is based on the average position.
Note2: When I say "higher ad position", I mean higher on the page. A "lower ad position" is a worse ad position on the page e.g. position 3 is higher than position 7.
Step 1) Buy Impression Data First
I build campaigns to count impressions first. This minimises ad spend, and also minimises time to market since you don't need to create anything other than a simple landing page to start with (and you could just point ads to the Yellow Pages if you don't want to create a page).
Your initial goal isn't really to get clicks, but to find out how often your ads are triggered, and then divide by your impression share to determine the total available number of searches in the given time period. (Note that you will need to have some spend with Google by way of clicks, or the campaign will eventually just die off - you see the impression share falling until these campaigns eventually stop running.)
You can often do this stage with minimal spend.
An example is me loading the "solicitors" seed in the UK, plus 40k locations. It ran the next day with a budget of £10. The ads were scheduled to run from 8am, and the budget was hit by 9am. The lowest reporting granularity in AdWords is an hour, so we didn't know whether the spend was within 10 minutes, or within 50-60 minutes. Anyway, in that time period and for our £10 spend, we bought 13 clicks, 700 impressions, and knew there were 4,000 searches we could have shown for (Impressions divided by Impression Share). Since we were bidding in +solicitors +location, and were one keyword per ad group, I was able to determine these metrics per location.
Step 2) Then Buy Click Data
If there is potential, I'll build new campaigns and run with higher bids to then start buying clicks. I'm then interested in the CPCs, average positions, and CTR and daily volumes at these higher ad positions. I'll also monitor impression share to see whether it's steady, or falling off a cliff again.
You don't get to see impression shares until a couple of days afterwards... which is annoying. You used to see it an hour afterwards, and you could often see new campaigns just dying within a few hours.
Your impression share should naturally be higher when you bid more, since you'll end up in a higher ad position where there are less advertisers able or willing to compete against you. There's usually only a few players competing for position 1, and a lot more players competing for position 6.
You might get some conversions at this stage, although I treat these as icing on the cake and "fluke" unless you hit a rich vein of conversions.
Something to note is that conversion rates can be higher at the start of a campaign and then drop off. My theory is that your ad is new, and appearing in front of people who've been looking for a while and are more frustrated than the people who will likely see your ad the week later. It's just a theory though, but I've seen this a few times so don't get carried away if you get a great conversion rate at the start.
You can push bids a little for selective hours or something, and determine what the CPCs and CTRs are for different ad positions. It's just indicative, since they will be different when your campaigns have been running for a while without change. I do this if my clients want to come up with some "what-if" scenarios... what if we had $1.00 CPC? What volume of clicks would we get for that CPC? What if we then doubled CPCs? What would our volume be for $2.00 CPC?
Step 3) Then Buy Conversion Data
Ok, so we've determined there are enough searches, and what clicks we can get for ballpark CPCs.
Now we want to find out our actual conversion rates, and our cost-per-lead or cost-per-sale. I'll just call these the cost-per-acquisition (CPA) for now.
This typically involves bidding to get a consistent stream of visitors through your funnel. Depending on your budget, and timescales, you can bid higher, or stay in a lower position. This will often depend on what you estimate your conversion rate and CPA will be. For lead-gen, I expect a 5% to 10% click-to-lead-conversion-rate if the funnels are tight enough (ST-AD-LP).. out of the box. If it's lower, then I'm a bit concerned and might tighten targeting (keywords etc).
I will often select a few of the search terms with the highest volume from step 2, so that the traffic is more consistent, and so that spend is not as spread across keywords. If you have budget constraints, then this is sensible. Also, it's always 80/20 anyway (often 99/1)... and a lot of your keywords aren't getting many impressions or clicks. Most won't be getting any impressions. Over a year they might pick some up, or get "activated" if there is an event that causes more searches to suddenly occur. So pick search terms/keywords with potential, and now use these.
You're now trying to get profitable. You need to determine daily and weekly (and maybe hourly):
The game now is to keep your CPC lower than your EPC, while still getting traffic.
You don't have to lose money now. You drop your CPCs until you're profitable. The only problem is if you drop them so far that you get no traffic anymore, and are no longer buying any data. Depending on your budget and timescales, you might push bids and knowingly lose money to buy data faster.
.......
You should now also watch this video.
EXAMPLE
You're spending $100 a day for 100 clicks, and making $50 revenue from 2 sales.
Metrics:
Cost = $100
Clicks = 100
Sales = 2
Revenue (Earnings) = $50
Calculated metrics:
Profit = -$50
cost-per-sale = $50
revenue-per-sale = $25
Or another way:
cost-per-click (CPC) = $1.00
earning-per-click (EPC) = $0.50
You're losing money. So drop bids until your CPC is $0.50. Then you're running breakeven. (This assumes that the traffic quality remains constant and doesn't get worse.)
You'll lose ad position, and your Impression Share and ad CTR will both drop.
I'd guess you'd end up with a quarter to an eighth the volume of inbound clicks. So about 12 to 25 clicks.
So you're breakeven, but have lost volume.
Now get into a continuous cycle of fixing leaks in your funnel, and increasing visitor life-time-value (LTV) so that your EPC increases. Each time you do that, you can increase your bids and drive more volume.
Apart from the initial budget to buy data and determine your EPC, there really is no reason to lose any money.
You just keep dropping bids till you're breakeven or are profitable.
You may have to drop them so low you get no volume... in which case you get no clicks, and therefore have no costs anyway.
How To Get Profitable Or Fail Fast With AdWords
Here's how I often determine whether a project, niche, seed, or keyword has legs, or needs a bullet.
Note1: I use blanket bid prices throughout the steps below, meaning the bid price within a campaign for each ad group or keyword is identical. I can then judge how much supply/competition there is based on the average position.
Note2: When I say "higher ad position", I mean higher on the page. A "lower ad position" is a worse ad position on the page e.g. position 3 is higher than position 7.
Step 1) Buy Impression Data First
I build campaigns to count impressions first. This minimises ad spend, and also minimises time to market since you don't need to create anything other than a simple landing page to start with (and you could just point ads to the Yellow Pages if you don't want to create a page).
Your initial goal isn't really to get clicks, but to find out how often your ads are triggered, and then divide by your impression share to determine the total available number of searches in the given time period. (Note that you will need to have some spend with Google by way of clicks, or the campaign will eventually just die off - you see the impression share falling until these campaigns eventually stop running.)
You can often do this stage with minimal spend.
An example is me loading the "solicitors" seed in the UK, plus 40k locations. It ran the next day with a budget of £10. The ads were scheduled to run from 8am, and the budget was hit by 9am. The lowest reporting granularity in AdWords is an hour, so we didn't know whether the spend was within 10 minutes, or within 50-60 minutes. Anyway, in that time period and for our £10 spend, we bought 13 clicks, 700 impressions, and knew there were 4,000 searches we could have shown for (Impressions divided by Impression Share). Since we were bidding in +solicitors +location, and were one keyword per ad group, I was able to determine these metrics per location.
Step 2) Then Buy Click Data
If there is potential, I'll build new campaigns and run with higher bids to then start buying clicks. I'm then interested in the CPCs, average positions, and CTR and daily volumes at these higher ad positions. I'll also monitor impression share to see whether it's steady, or falling off a cliff again.
You don't get to see impression shares until a couple of days afterwards... which is annoying. You used to see it an hour afterwards, and you could often see new campaigns just dying within a few hours.
Your impression share should naturally be higher when you bid more, since you'll end up in a higher ad position where there are less advertisers able or willing to compete against you. There's usually only a few players competing for position 1, and a lot more players competing for position 6.
You might get some conversions at this stage, although I treat these as icing on the cake and "fluke" unless you hit a rich vein of conversions.
Something to note is that conversion rates can be higher at the start of a campaign and then drop off. My theory is that your ad is new, and appearing in front of people who've been looking for a while and are more frustrated than the people who will likely see your ad the week later. It's just a theory though, but I've seen this a few times so don't get carried away if you get a great conversion rate at the start.
You can push bids a little for selective hours or something, and determine what the CPCs and CTRs are for different ad positions. It's just indicative, since they will be different when your campaigns have been running for a while without change. I do this if my clients want to come up with some "what-if" scenarios... what if we had $1.00 CPC? What volume of clicks would we get for that CPC? What if we then doubled CPCs? What would our volume be for $2.00 CPC?
Step 3) Then Buy Conversion Data
Ok, so we've determined there are enough searches, and what clicks we can get for ballpark CPCs.
Now we want to find out our actual conversion rates, and our cost-per-lead or cost-per-sale. I'll just call these the cost-per-acquisition (CPA) for now.
This typically involves bidding to get a consistent stream of visitors through your funnel. Depending on your budget, and timescales, you can bid higher, or stay in a lower position. This will often depend on what you estimate your conversion rate and CPA will be. For lead-gen, I expect a 5% to 10% click-to-lead-conversion-rate if the funnels are tight enough (ST-AD-LP).. out of the box. If it's lower, then I'm a bit concerned and might tighten targeting (keywords etc).
I will often select a few of the search terms with the highest volume from step 2, so that the traffic is more consistent, and so that spend is not as spread across keywords. If you have budget constraints, then this is sensible. Also, it's always 80/20 anyway (often 99/1)... and a lot of your keywords aren't getting many impressions or clicks. Most won't be getting any impressions. Over a year they might pick some up, or get "activated" if there is an event that causes more searches to suddenly occur. So pick search terms/keywords with potential, and now use these.
You're now trying to get profitable. You need to determine daily and weekly (and maybe hourly):
- Costs
- Clicks
- Conversions
- Revenue (Earnings)
The game now is to keep your CPC lower than your EPC, while still getting traffic.
You don't have to lose money now. You drop your CPCs until you're profitable. The only problem is if you drop them so far that you get no traffic anymore, and are no longer buying any data. Depending on your budget and timescales, you might push bids and knowingly lose money to buy data faster.
.......
You should now also watch this video.
EXAMPLE
You're spending $100 a day for 100 clicks, and making $50 revenue from 2 sales.
Metrics:
Cost = $100
Clicks = 100
Sales = 2
Revenue (Earnings) = $50
Calculated metrics:
Profit = -$50
cost-per-sale = $50
revenue-per-sale = $25
Or another way:
cost-per-click (CPC) = $1.00
earning-per-click (EPC) = $0.50
You're losing money. So drop bids until your CPC is $0.50. Then you're running breakeven. (This assumes that the traffic quality remains constant and doesn't get worse.)
You'll lose ad position, and your Impression Share and ad CTR will both drop.
I'd guess you'd end up with a quarter to an eighth the volume of inbound clicks. So about 12 to 25 clicks.
So you're breakeven, but have lost volume.
Now get into a continuous cycle of fixing leaks in your funnel, and increasing visitor life-time-value (LTV) so that your EPC increases. Each time you do that, you can increase your bids and drive more volume.
Apart from the initial budget to buy data and determine your EPC, there really is no reason to lose any money.
You just keep dropping bids till you're breakeven or are profitable.
You may have to drop them so low you get no volume... in which case you get no clicks, and therefore have no costs anyway.
Don't like ads? Remove them while supporting the forum:
Subscribe to Fastlane Insiders.
