Discuss Home · Bootstrapped Podcast · Scribbleton Personal Wiki · HelpSpot Customer Service Software · Thermostat NPS

The Black Art of AdWords


So, I’ve been trying to drive traffic to the SmallSpec page in order to collect e-mail addresses. To do this, I’ve been using AdWords.

To date, out of 35,000 impressions there have been exactly 115 clicks. That’s a click-through rate of…well, it’s not very good.

I’m not sure what I’m doing wrong, but I’m pretty well convinced that I am doing something wrong. Perhaps my keywords are garbage. Perhaps it’s too early for AdWords, and I need to be doing more of the non-scalable stuff like blogging, talking to people, etc in order to drive traffic. But at the same time, I can’t help but wonder - if I can’t get sufficient AdWords traffic to build up a mailing list now, how will I ever get sufficient AdWords traffic to make the app viable?

Can anyone suggest resources that can help me learn how to use AdWords more effectively and identify what I’m doing wrong?


Where’s the @patio11 bat-signal?


Oooooh, this is in his wheelhouse, isn’t it?


I just bought the Ultimate Guide to Google Adwords which was recommended by Rob Walling’s Micropreneur Academy. You can probably find at least a few new tricks in there to help improve your results. I’m just getting started, but the book seems solid.

For example, according to the book:

Under performing ads within the same Ad group will bring down the group as a whole. The recommendation of the book is to “Peel and Stick” the better performing ads into their own group - i.e. the ads that have higher click-through rates.

Also, ads should be grouped into specific groups based on keywords that go together. It is now considered bad practice to fill up your ad group with a ton of varied keywords.



I’ll check out that book, thanks!


I am by no means an AdWords expert, but from what I’ve picked-up the solution to low CTR can be summed up in one word: relevance.

Two factors that have the greatest impact on CTR are keywords and ad copy.

Keywords: Crawl inside the head of your customer, identify their need, and figure out what they would type into google to find the solution. Playing with Google Suggest might help with this. The more relevant the keywords are to the problem/solution you’re targeting, the higher your CTR will be.

Ad Copy: Your ad copy should speak directly to the user’s need and offer a solution. Ad Groups are what allow you to make the ad copy relevant. The keywords you put into an Ad Group should be closely related to the ad copy in that group. Increase the affinity between your keywords and ad copy and you will increase your CTR.

Just my 2 cents.

This is a really good talk by Dave Collins at MicroConf 2012: http://vimeo.com/51187193


It’s very hard to offer actionable advice without having seen the specific campaign at hand. But without loosing generality, I strongly suggest working on preserving smell.

Effectively, each search carries a certain intention with it. Specifically eg. people typing in “wedding dress” is a very strong evidence of the user being 25-40 female, in a relationship that is about to go Serious, with a willingness and ability to command a large budget, and the intention of buying the prettiest dress the internet has to offer. The ads, and sites which are the most responsive to these circumstances, and intentions will win both in SEO, and paid advertising.

What struck me most about your post, is the lack of hypotheses behind the thinking processes of the users. The rationale behind “talking to customers” is exactly to build such hypotheses, so it can inform the rest of the channel building process. Specifically, you might find, that neither engineers nor managers are interested in “functional specifications” -rather, they might look for ways to “improve communication”, or “remote collaboration”. You might find brute-forcing the infinitespace of tokens a wee bit harder, than having a 15 minute talk with a relevant decision maker.


One of the biggest mistakes I see with AdWords (and Google more or less encourages you to do so) is starting out with broad match keywords. IE you are selling “Help Desk Software” and bid on the broad match keyword “help desk”. This is bad for two reasons, one people googling “I need help assembling my desk” will see your ad. If you are lucky they won’t click your ad, but more often than not people don’t read and just click everything. If you are doing any broad match you’ll almost certainly need to waste a bunch of money just to setup a huge negative keyword list. You’ll also get a low click through rate and a bad quality score.


Eeeeeeh. I clearly don’t understand what I’m doing. I don’t completely understand why certain keywords I’ve chosen get clicks while very similar keywords don’t. It’s aggravating. And it seems like the rules are constantly changing.

I’m probably going to have to hire out to get help with this. And that frustrates me, because I really do want to master this tool. But there’s just something about it that I don’t “get” yet.

I’ve started reading “Ultimate Guide to Google AdWords” but no light bulb moments so far.

What struck me most about your post, is the lack of hypotheses behind
the thinking processes of the users.

Well…I’m building a functional spec tool If you want to build functional specs, my tool may be useful. There’s no hypothesis to be had there, it’s the simplest scenario in the world.

The rationale behind “talking to customers” is exactly to build such
hypotheses, so it can inform the rest of the channel building process.

Don’t get me wrong, I’ve talked to potential customers in the form of fellow freelancers. The general takeaway so far is " When I want to build functional specs, that tool may be useful". I’m not doing anything complicated here.

"Specifically, you might find, that neither engineers nor managers are
interested in “functional specifications” -rather, they might look for
ways to “improve communication”, or “remote collaboration”.

See, this may be why I don’t “get” AdWords, but that makes no sense to me. None. If you need to create functional specs and don’t know how, or don’t have a good tool, Googling “functional specification” makes way more sense than Googling “improve communication”, which seems really broad, even more so than “functional specification”.

Is my thinking on this subject just broken? Am I hopelessly locked into “engineer-itis” when it comes to understanding marketing?


OK, I’m reading the recommended book. It’s already made a big difference.

After completely deleting my old AdWords campaign I created a new one using the guidelines from the book. The difference has been huge, even though the volume is small since the new campaign has only been running for 2 or 3 days.

The old campaign:
35,000 impressions
115 clicks
1 sign-up

The new campaign:
2,350 impressions
19 clicks
10 sign-ups (!)

The over all click-through rate (CTR) is still nothing to write home about, but it’s miles better than the old one. Plus, after reading the bit about the “Peel & Stick” strategy, I’m getting ready to pull some of the keywords out into their own Ad Group, which should improve the CTR even more.

So…even for a clueless soul like myself, for whom AdWords does not come naturally, a bit of expert guidance can apparently go a long way.



I have a question for those of you who are more experienced with AdWords than I am.

So, I’m using the “Peel & Stick” strategy for gaining a higher average CTR. Marshall’s book says to create a new ad group and a new ad for each keyword, using the keyword as the headline.

In my case, the keywords were all too long to use as headlines, so I tried to instead incorporate them into the ad itself, which (in my opinion) made for slapped-together, less-compelling ads. I guess the CTR over the next few days will tell me if I’m right or not about the ads being lower-quality. But the first week of results seem to be telling - every one of the keywords I’ve peeled out of the generic ad group have performed worse on their own.

So, my question is – has anyone else run into the problem of too-long-for-headlines keywords while using the “Peel & Stick” strategy? If so, what did you do? And did you find that - regardless of whether the keywords were too long for headlines or not - there was a performance drop in terms of CTR, instead of the expected boost?

I’m pretty sure I did something wrong here, but I’m not sure what.