Heroview - Unveiling Display Network and ACE Tactics with Melissa Mackey

For today’s #heroview, it was our pleasure to sit down with PPC guru Melissa Mackey (@Mel66) to talk about two tricky topics – AdWords Campaign Experiments (ACE) and the Display Network.  Here is the transcription of the discussion!

Thank you to everyone who participated in this month’s Heroview – real-time discussions with PPC industry experts via Twitter.  Stay tuned for next month’s Heroview!

June 17, 2011

PPC Hero: Welcome to Heroview! To get started, tell me a little about yourself. How long have you been working with PPC?

Melissa: I’ve been doing PPC since 2002, when Adwords first launched; first in house and now with @fluencymedia. I’m “old school PPC.” 🙂

PPCHero: That’s great – agency and internal experience really makes a great PPCer! So today we’re going to talk about two advanced areas of PPC: Adwords Campaign Experiments (ACE), and the display network.

Melissa: 2 of my favorite topics!

PPC Hero: When do you use ACE? Is it better for campaigns that are fairly stable, or for campaigns that need a lot of work?

Melissa: I like to use ACE for both stable campaigns (to test new things) and troubled campaigns (to improve performance). That said, it’s most leveraged for campaigns that need quick triage, but also need good data to prove results.

PPC Hero: Could you give us a quick example of something you’d used ACE for in a stable campaign?

Melissa: I like to use ACE to add “crazy” keywords to a stable campaign, just to see if they might work. I’ll set the % low and watch it. You never know what might work for your account!

PPC Hero: Interesting! How about for a campaign that’s troubled?

Melissa: Lately, I’ve been using it for bid tests: trying lower or higher bids to see what happens with conversion rate. This can often show that lower bids don’t affect conversion rate, and therefore lead to a better cost/conv.

PPC Hero: How about for testing ad text? Do you recommend using ACE or putting ads on rotate for A/B testing?

Melissa: I usually do straight A/B ad tests, or A/A/A/A/B if I have a strong hero. I like to control things vs. letting G do it. But ACE also works well for ad testing, especially when you’re using CPA bids – you often get better data. And of course be sure to put ads on Rotate, always. 🙂

PPC Hero: Is there anything that ACE doesn’t test that you would like it to, such as demographic targeting and dayparting?

Melissa: I’d love to test campaign settings, like dayparting, networks, ad delivery, etc. but it’s not possible now. I’d especially like to experiment with the Search network on vs. off. Would be very insightful. Dayparting would be great too – to prove assumptions about what days & times work rather than guessing.

PPC Hero: That would be nice! Hopefully they’ll get on that soon.

Melissa: Agreed!

PPC Hero: Finally, what would be your advice to someone who has never used ACE before?

Melissa: Read the documentation carefully! It’s easy to misunderstand “experiment only” and “control only” and you can end up with lots of your keywords paused. Don’t ask how I know. 🙂  Anything that’s “control only” will go away when you “roll out changes fully.” Forewarned is forearmed. 🙂

PPC Hero: Great advice. Okay, let’s move on to the display network! In this post (httPPC Hero://ppche.ro/j65urc), you say that display isn’t for everyone. Can you elaborate on who shouldn’t use it?

Melissa: If you’re at all risk-tolerant, display may not be for you. It can work great, but does have risks. I meant “risk INtolerant.” 🙂 (darn the fast fingers!)

PPC Hero: Ha ha, yes, important distinction 😀

Melissa: If you don’t want to or can’t take the risk of getting a lot of non-converting traffic, you may not want to use Display. It also needs to be managed separate from search, so if you’re really busy, display may not be for you.

PPC Hero: Do you ever recommend creating display campaigns with only image/media ads, and no keywords?

Melissa: Sure! It can be a great way to cast a wider net for your campaign and get lots of branding impressions w/o the limitations that keywords can impose. And image ads can include rich media & animation, which is cool too.

PPC Hero: Interesting. How about exclusions? You suggest using exclusions in the display network, but do you also use managed placements to keep your traffic qualified?

Melissa: Absolutely. I often start without managed placements, then add the ones that perform well to managed placements. Or I’ll move those placements to their own campaign, and then exclude them on the original to avoid confusing G. 🙂

PPC Hero: Good thinking!

Melissa: I find it easier to learn which placements perform well, as opposed to guessing by starting out that way.

PPC Hero: Getting quickly back to media: do you have any advice for PPCers who want to do image ads but don’t have graphic designers?

Melissa: 3 words: Display Ad Builder httPPC Hero://bit.ly/iM7DKK. You can create your own image ads without a designer. Awesome! If you can get logos or images from your art dept. (or your client), even better. The Builder is easy to use and is a great tool for creating lots of different display ads to test.

PPC Hero: Very good point. Alright, let’s get onto our last question! You have a blog coming out soon that will share display network resources. Can we have a sneak peek at your favorites?

Melissa: Sure! Check out the post: httPPC Hero://bit.ly/fCYumh. One of the resources mentioned is the free ebook Customers Now by @szetela. Full of good tips. httPPC Hero://bit.ly/1UpdGU. I regularly refer to that book for new ideas, even though I use display a lot.

PPC Hero: Great! Can’t wait to read the full article. Well, that’s all we have for today! Does anyone have any questions for Melissa?

-@JARooney8423: have you tried the topics tab for the display network yet? what do you think?

Melissa: @JARooney8423 Good question! Yes, we were part of the beta for that. I love it – we get really good results there.

@Jessica_Cates: Is there any area of the DN that you find most people aren’t utilizing enough? Ex. topics, remarketing, etc.

Melissa: @Jessica_Cates Remarketing, for sure – even I am barely using that yet. 🙂 I think it’s tough to use b/c of the tagging, but is really* leveraged and should be a higher priority for advertisers than it is.

-@Hoffman8: To follow up on @Jessica_Cates question, which type is your favorite? For both leadgen & Ecommerce 🙂

Melissa: @Hoffman8 Tough question! I’d say it depends: display can be hard to do for ecommerce, although it can be done for sure.

@Jessica_Cates:   I can see that. I’m using it here and there with great results, it makes me sad if I don’t have access to add tags 🙁 Thx.

Melissa: @Hoffman8 I’m becoming a huge fan of Topics, as long as there are available topics that are relevant to your campaign.

@JARooney8423: that can be a problem. do you know how often they add new topics (categories and subcategories)?

Melissa: @Hoffman8 For leadgen, I like text ads best; especially for B to B. I think B to B searchers are immune to image ads. 🙂

Melissa: @JARooney8423 Another great question! I don’t know – but will try to find out!

@LisaSanner: have you had success with category targeting overlayed with placement targeting for DN?

Melissa: @LisaSanner Not really – I’ve found that really limits the available impressions, so we don’t get much traffic. In other words, lots of work for not much payback. 🙂

Melissa: You folks ask good questions!! 🙂

@Hoffman8: I love Topics, too. I’ve had better luck using KW+Topics than just Topics, any thoughts?

Melissa: @Hoffman8 Again I think it depends. Topics alone is too broad for some advertisers, for sure. Also depends on adv. goals. I.e. strict direct marketing goals vs. high impression & traffic goals, price of product/service, etc.

-@Hoffman8: Makes sense! Thanks 🙂

PPC Hero: Alright, time to wrap up today’s #heroview. Thank you @Mel66 for such great, insightful answers!

Melissa: My pleasure! This was fun!

PPC Hero: You can find the transcription of today’s #heroview on our blog, ppchero.com (the link will be tweeted). Thanks again to everyone!