This is the second in a series of interviews we're conducting with AdWords advertisers who got unusually high scores using our AdWords Performance Grader . We're reaching out to high scorers to get AdWords help  and find out what strategies contribute to their strong AdWords performance. For more in this series, see:
Today’s interview is with Stefan van Vliet, co-owner of Dutch agency Compass Online Marketing .
Tell us a bit about yourself. How long have you been using AdWords? Are you an Agency or an Advertiser? What is your primary goal for AdWords marketing?
I have been working as an SEO consultant for a small Dutch web marketing agency  for some years. Last February I launched a webstore, Feestkleding 365 , selling fancy dress clothing. I was determined to launch before Carnaval (a big happening here in Holland) in order to tap into the massive demand. As SEO traffic is non-existent with a new site, I decided to use AdWords as our only sales channel. With some knowledge from a book and a few blogs but no experience, I put together an AdWords campaign. Sales went through the roof, but so did our AdWords costs. I chose to set the budget to virtually limitless and tried to optimize our campaign while running, but just couldn’t keep up (and honestly didn’t know what I was doing). So only 1 week and almost €10k in losses later I pulled the plug. So far for an expensive lesson in AdWords management: always test before you allocate serious budgets.
Can you describe your AdWords management strategy? How do you set your campaign objectives, and how do you know what’s realistic or not?
It took a while to find a strategy that was working for us. The thing is that our product assortment is quite large: we offer costumes and accessories for a big variety of themes. At the moment we are only targeting approximately 10% of those themes, and we gradually add campaigns for new themes as we go. Our strategy is something like this:
- Get sufficient data by bidding high and getting in spot #2-3 (above organic results).
- Keep track of conversions, cost per conversion and value per conversion (using conversion tracking with dynamic value tracking).
- Adjust bidding to meet a cost per conversion which is roughly 50% of the value per conversion.
- Optimize ads and try to improve conversion rates .
What makes this process quite difficult, is that although demand for fancy clothing overall is huge, search volume per theme can be quite low, which means that it can take quite some time to gather statistically significant data.
Any advice or tips for AdWords marketers that didn’t score as well as you?
Here’s my advice:
1. Leverage negative keyword lists. This helps us to efficiently manage our 20+ campaigns effectively. For example: we do not target keywords containing city names, as conversion rate is very low (apparently people looking for fancy clothing in a certain city do not wish to buy online). So we added 1,200 city names as negative keywords . If we didn’t use negative keyword lists, certain negative keywords, which only apply to 1 campaign, would easily get lost in the huge amount of other negative keywords. As some negatives are only added temporarily (for example certain products which are sold out but will be back in stock in a few months), it would be very hard to spot and remove those negatives.
2. Split campaigns on a low level. We decided to create a campaign for every theme, which allows us to add sitelinks related to that theme, and to easily segment data based on themes. Within the theme campaign, we add ad groups targeting keywords like [theme] + [clothing]. For example, within the Halloween theme, we have a separate ad group for Halloween clothing, Halloween costume, Halloween Fancy Dress etc. This improves CTR  dramatically. Of course, this would only work if sufficient search volume exists on either of these ad group level keywords. At first, we even added ad groups targeting keywords like [theme] + [clothing] + [gender]. So we had ad groups for Halloween costumes for men, Halloween costumes for women, Halloween costumes for boys etc. But we soon found out that those Ad groups got ignored by Google due to low search volume . So these groups basically just cluttered our account and we decided to delete them.
3. Leverage advanced CPC bidding in Google AdWords Editor . If we add a new campaign, we want to gather data first, so we aim for spot 2. Sorting on current position and bulk editing CPC’s by +/- 10% is a fast and easy way to get your ad on that position until you have sufficient data. Be aware: choose a short timeframe when sorting on position (i.e. 1 day), as you don’t want to know the average position but your actual one. Also double-check position for keywords in the ad group: sometimes they vary greatly, which means you would have to split them up into different groups in order to effectively manage their CPC bids.
4. Use the comments function within AdWords Editor  to make notes of what you do. Note date, current position, CTR, etc. and of course actions taken. This will keep things organized and help you see the effect of your actions immediately.