This holiday season has seen a phenomenal increase in people looking to buy a Nintendo Wii. Like any popular niche, it brings hordes of people trying to make a buck. Because it is so highly competitive (can we say “ringtones”?), it weeds out the newbies, until finally the battle is waged between those who don’t need to make a profit (e.g. places like Best Buy and eBay who don’t need to profit on every single transaction) and the really smart internet marketer.
So a little history of what I’m doing:
At $100+ per sale, you might think that I should have just stocked up on Wii’s and sold them for massive profit. But how many could I buy (even if I could get my hands on some)? 10? So I’d invest $2490 (assume $249 per Wii) to make $1000. 40% profit is not bad for an online store, but I’m looking for something better. And sooner or later (when Nintendo brings in more stock) the price will drop, and I would be stuck with Wii’s (that I might have bought at a higher price) that I can’t offload profitably.
Now onto the Adword details. This post isn’t about all the details for setting up an Adwords campaign. It specifically focuses on adcopy (creating ads for an Adwords campaign). There are so many posts out there telling you to do A-B testing of your ads, but hardly anyone explains what that means, and the few that do never give real-life examples. So I thought fellow Internet Marketers out there might benefit from the chronological history of my ad testing: what ads I created (and why!), what I learnt from the results, and how small changes can have a big impact.
Before I start, some background:
Before I write any ads, I search with that keyword, to see what adcopy my competitors are using, and then create a better selling proposition. When I did this research for this particular keyword, I found that:
So I start my campaign with these 2 ads:
Nintendo Wii ™ - In Stock Nintendo Wii Systems, Games & More In Stock & Ready To Ship Today!
Nintendo Wii ™ In Stock Nintendo Wii Systems, Games & More In Stock & Ready To Ship Today!
I decided to use ™ in my adcopy, because Wii is a legitimate trademark and I’ve learnt from past campaigns that it is a good differentiator (it makes my ads stand out, so an automatic CTR boost). I also used “In Stock & Ready To Ship” because this was the big problem currently (stores like Best Buy advertising Wii, but not actually having it in stock).
In this initial test, the only difference is the “-” in the heading. The trick to A-B testing is to only have one difference, so that you can easily identify what works or not. The other (more important) trick to any testing is to wait to get enough impressions/clicks to form a valid decision (what I call a “statistically significant” sample). After obtaining at least 30 clicks for BOTH ads, I found that the first ad had a CTR of 1.77% and the second ad had a CTR of 1.98%.
When I start a new A-B test, I don’t over-analyze about why one ad worked or not. I don’t want my campaign to suffer from low CTR for too long, so over-thinking is not a good idea. After all the A-B testing I’ve done, I’ve come to the conclusion that there is mostly no logic to what works or not. Just kill the one that doesn’t work, and try the next trick on your list.
So I had to replace the first ad, with a version that was based-on but better than the second ad. This is what I came up with:
Nintendo Wii ™ In Stock Buy Wii Systems, Games & More In Stock & Ready To Ship Today!
Here I changed the 1st line. Because the keyword I was writing this ad for had “buy” in it, I tried to put that keyword in the ad so that I’d get more of my ad highlighted. I had avoided using this on the initial adcopies because all my competitors had “buy” in their ad, and the objective is to make my ad stand out (sometimes not getting highlighted is a better technique). As I had suspected, getting “buy” highlighted was not useful. This new ad had a CTR of 1.63% (worse than both the original ads) so this had to be replaced urgently with the next version:
Nintendo Wii ™ In Stock Nintendo Wii Systems, Games & More In Stock & Ready To Ship Now!
Here I replaced the “Today” with “Now”. With a huge CTR jump to 2.23% (and now my #1 perfoming ad), this was clearly a good move. Once a CTR exceeds 2%, I try to slow down and think more about why things work or don’t work. I guess in the urgency of the holidays, “Now” resonated better. Might be only something here…
At this time, I still had some keywords/ads that were <2% so I turned my attention to them. I have found that a CTR >2% is good: over time my aCPC drop and/or my ad position increase (by itself).
In case you didn’t notice the really good piece of Adwords advice here: CTR of 2% or better is what you need to aim for to naturally lower your cost over time. The magic number actually varies (usually around 1.5%) but I’ve haven’t yet found a keyword that needed more than 2%.
Anyway, in the span of 3 days that I was focused elsewhere, I found my CTR for this good ad had dropped slowly until it was <2%. My ad position had stayed constant, so something must have been up. I did another Google search and noticed that my competitors had changed! That’s one thing about a popular niche… you have to stay on top of things every day or even a successful and profitable campaign can get buried overnight. What I discovered was that >50% of the ads on the first page were now affiliate sites (other BANS and Amazon stores). And nearly 75% of the ads (all the new competitors and some of the original ones) had copied my “In Stock” and “Ships Now” adcopy! I’m not too worried about these new competitors because:
But I didn’t like it because:
So once I had gotten my other ads up to 2%, I returned my attention to improving the CTR. Just to be clear, I always work on improving my ads (because ads will lose their freshness over time, it’s really the easiest/simplest way to get more clicks, and the more above-average my CTR is the faster that Google will decrease my aCPC). So I came up with another version to compete with my #1 ad:
1000s of Wii In Stock Now Buy Nintendo Wii Before Sold Out In Stock & Ready To Ship Now!
I came up with this adcopy to specifically combat the fact that everyone had copied my previous ad style:
This ad paid off big time, the CTR surging from 2%, to 3.14%. With little effort (it took me less than one minute of thinking), I increased my click-thru’s by 50%! Not content, I wanted to put more distance between myself and the copy-cats ads… I came up with this:
1000s of Wii In Stock Now Buy A Nintendo Wii Before Sold Out In Stock & Ready To Ship Now!
The only difference is that I added a “A” before “Nintendo”. I decided to do this after reviewing Google’s Search Query Report (which shows what people are searching for when your ad is displayed). When I first started my campaign, people were searching for “find wii online”. Whereas now, there were more searches like “find a wii online”. I figured that throwing an “A” in the ad was worth trying. This further increased my CTR to 3.31%.
Coming back to my original thought that “Now” resonated with the searchers, I thought I’d try moving the word around:
1000s of Wii Now In Stock Buy A Nintendo Wii Before Sold Out In Stock & Ready To Ship Now!
Wham! By moving the “Now” (in the heading) before “In Stock”, my CTR shot up to 4.18%. With a simple repositioning, I increased my CTR nearly 1%. I’m a firm believer of “peripheral vision” for adcopy. When people see “Wii” bolded in the heading, they’re also sometimes seeing the word(s) next to it (which would be “Now” instead of “In”). As a result, people who see the “Wii” also absorb the “Now”, resulting in a greater urgency to click on my ad. Is this pop-psychology true? I don’t know, but I’ve found this “peripheral vision” works in most of my ads. But I recommend you don’t use the “peripheral vision” technique until you’ve locked down a high performing ad (because it doesn’t normally give such a big improvement, and while you you waste time tweaking a bad ad, you should be focusing on getting past the 2% CTR target).
And that’s where this story will end for now… My campaign is still rocking along, so we’ll see how it goes into 2008. I’m sure the search popularity for Wii’s will drop when there is more availability, but hopefully by then, my aCPC will be so low that I can continue making a tidy profit in this niche.
Good stuff Ian
Great post Ian! It really helped me get a bit of insight as to how you do some of your stuff! I really appreciate it!
Wow, VERY fascinating! Hope you are well. -Trent
[...] so I’m suggesting that you all read it now! Its not often you see someone talk about Split Ad Testing and actually show some of their own examples. It’s written towards Google Adwords but just [...]
This is a very powerful and informative post. You really elaborate on not only A/B testing but also tell people how you structure your campaigns.
fascinating, and thx for the insight
Wow, this article is amazing. I have never seen an article about split testing that is this specific. Usually they get lost in trying to give general guidelines and don’t end up teaching this much. You have really given great insights.
Also, although it looked bad there for a moment when everyone copied your ad, it may have turned out in your favor with your next moves because you had a chance to stand out from a more uniform crowd. Although in a competitive field, especially riding a short term buying frenzy, you have to keep tweaking every day, so no gain is something to rest on, except that as your skills grow, tweaking and adjusting gets easier and less time consuming.
Smith