SLC|DMC took a different approach this year, but it was still filled with great insights and important discussions from industry authorities and insiders. Last year, we fit a full range of internet marketing topics into a single day, punctuated by keynote address by the likes of Lee Odden and Cyrus Shepard. This year, we split the conference into two days, and one would address paid search marketing while the other would look at organic SEO.
The organizers of the event said that this was done deliberately to give more people a chance to start expanding their horizons and really get some exposure to a different side of the industry. So for many people, this dive into the deep end of paid search marketing was completely new – but each of the speakers brought something insightful and interesting to the table.
Combatting Rising Cost Per Acquisition with Audience Targeting | Brad Geddes
Brad Geddes, author, speaker, and marketing consultant, opened this year’s SLC|DMC with a keynote discussion of how costs per acquisition (CPAs) keep going up, year over year, and what, exactly, we can do about it. His claim, the one that should be the most important to marketers and paid search managers, is that you could achieve some major drops in CPAs in just 44 minutes of work.
How, exactly, can this be done? Is the CPA so fluid that it really only takes a few changes to keep it down? And if it’s really so easy, wouldn’t everyone be doing it?
In order to find out, we started by looking at the new reality of consumers online.
Consumers, Geddes said, now visit websites at least 6x on average in the purchase process. (You’ll note as we go through all of these presentations that the consumer buying cycle and the amount of research they do before purchasing is a reoccurring theme.) This, he said, is an important element of a PPC marketing strategy and can help improve the average conversion rate (which can be said to be somewhere around 2%).
Consumers spend more time with information than ever before, simply because they have more access than ever before. You need to consider this in your ads and how you measure your conversion rates.
Geddes brought up the notion of Relative Conversion Rate vs. Absolute Conversion Rate. This is important because calculating conversion rate off the number of unique visitors is probably more helpful and accurate than calculating it off of all the visitors.
So, for PPC practitioners, you need to start looking at more specific numbers when you calculate things like the conversion rate. But it’s more than just unique visitors. You can then go another step and start targeting different audiences. (Another theme that you’re going to see revisited throughout the different presentations.)
There are, he says, several ways that you can use paid ads and remarketing to make sure that you’re giving your consumers every chance to choose you over the competition. First, you need to create remarketing lists for search ads – this is to target those people who started the buying or investigation process on your website and then left.
This kind of targeting is effective when you’re looking at the lifetime value of the customer. This will help you follow through with each one.
The longer the customer journey, he said, the more you need remarketing. Consider the target audience in this process. If you are working on a travel website or something else that involves a major purchase, they will almost never make a purchase on their first visit. So, it’s important to stick to it and make sure you’re using the right data to make the right decisions.
Customer Match – Remarketing Based on CRM Data
Customer match and remarketing can both benefit your campaign in different ways.
- Online data
- Must tag website
Customer Match, on the other hand, offers a chance to get:
- Offline data
- CRM data
- Can be “tagless” remarketing that does not require website codes.
Brad then went on to discuss how dynamic search ads can also figure into a campaign like this because it will actually leverage organic crawl data to ensure a more relevant audience.
What Can You Do with Audience Data?
So after talking about the importance of properly segmenting your audience by data, what else can you do with this to help you get that CPA back down to a reasonable rate. With the right data, you can:
- Change your bids
- Add and target negative keyword lists
- change your ads to mirror the sales funnel
- Expand keywords
- Change keywords
Audience data will also let you start to properly segment your lists which, in turn, will give you more insights into consumer behavior. Brad suggested a few ways to segment your lists. You could, for example, segment it into employees (for a negative list), blog readers, site visitors, those who downloaded something, and those who asked for a free trial. This will help you specify whether you should target them with remarketing or employ other methods.
In the end, Brad said that because he looks at behavior and properly segments his audience, he can leverage that information to lower the PPC CPA in just 44 minutes per campaign.
“Audiences,” he says, is a term that should be a part of all search marketing. This is what will help you make sure that someone will come back the average 6x it takes to finally convert.
Effective Storytelling with Data Analytics | Luciano Pesci
The theme of this presentation was to think big, mine deep, and explain simply. Luciano Pesci has completed hundreds of primary research and data mining projects for tech focused startups, small businesses, and nonprofits, and he brought that experience here to show how we can use the right kind of data to improve PPC campaigns.
The basis for this presentation was his claim that “statistics is the worst-taught subject in academia,” and that this is in spite of the fact that data skills, he says, are as “important as typing kills in the early 80s.” (That’s more of a paraphrase than a quote, but that was definitely the intent of the message.)
So why is it so important to understand data these days?
- It’s never been easier to gather data
- Data is the competitive decider right now (which is how Google and Amazon crush their competition)
- Data doesn’t prove It never does. But it can support almost everything.
- Data should be able to complement your gut intuition – you don’t need to blindly follow statistics
Perhaps the most important statistic in this discussion of statistics is that 80% of data projects are failing right now because their analytics lack meaning and context.
So what does storytelling have to do with data analytics? In Luciano’s words, if you can’t explain the data to an 8-year-old, it’s not ready for upper management. But everyone understands a story, so then we got into how we might use the data to become good storytellers.
In order to tell the best story, you need to go through three steps:
- Think broadly – Capture all the relevant data you can
- Mine deeply – Identify what it is you want to analyze
- Explain simply – Translate the numbers into a real story
As an example, he took us through one of the most popular YouTube channels currently producing content: Epic Rap Battles of History. His goal was to see if he could tell their full story through data. After all, this is a channel with only 64 episodes but more than 3 billion total views.
We’ll skip over all the details of the story, suffice it to say that by mining deeply into even the publicly available data led to some interesting conclusions. This would be the “think broadly” step, because it involved looking into all the available information.
In any analytics project, he said, you need to define a very clear target. Only then will you be able to see what’s driving engagement and whether or not the product/service/content is living up to the brand promise.
In this case, he was able to look at the data and create a set of new variables – likes, views, comments, language used, etc. This would fall into the “mining deeply” element because it is where he identified what, exactly, he wanted to learn.
Finally, we talked about how you can explain the data simply. We had to ask the question of what’s driving engagement and what can be said about what we observed.
Once we can tell that story, we can start to answer other questions, like whether (in this case) we should start to advertise the most popular videos to get even more engagement, or the ones that get the most likes or the most comments.
The data was able to show that Epic Rap Battles of History is well aware of how their audience wants to engage with their brand, and they are living up to their brand promise. So the real question here is not who would win in a rap battle between Bill Gates and Steve Jobs, but whether you can tell the same kind of story with your own analytics.
Mastering the Tools and Tactics of Google Display | Susan Wenograd
Because we don’t call them “banner ads” anymore, Susan Wenograd said, we are going to spend this time talking about display advertising. In order to get the most out of this form of online advertisement, she recommend two different types of ad targeting: by site and by people.
Site Focus Targeting
When your display ads have a site focus, you can reach specific audiences through:
- Topical targeting – Showing ads only on websites with relevant topics
- Keyword targeting – Stick to around 5 keywords
- Placement targeting – Define specific sites, but be aware that you can’t specify where on that site, so you could end up on irrelevant stuff
- Audience Targeting – Back to defined audiences again. This can be in-market people or affinity people (the fan boys). She also discussed how you can use layering or overlapping tactics to get to custom audiences.
Optimizing Display Ads
Susan talked about the number of times she’s heard people claim that display advertising no longer works. However, just like many online advertising techniques, it works just fine, but not alone. You have to optimize it to get the most benefit. In this case, she gave three things to focus on.
- Ditch bad placements – You should definitely use placement targeting.
- Impression Caps – Look at conversions per impressions and set your cap where you know there’s a diminishing return.
- Negative Keywords and Topic categories – Know what to exclude. Get rid of words and placements that don’t convert and always exclude “Sensitive Content” in the categories.
People Focused Targeting
Focusing on people requires much of the same tactics discussed in Brad Geddes’ presentation. It all comes down to creating an audience based on behavior, but in this case Susan also recommended segmenting by timeframe, if necessary.
Time-based marketing says that it’s just as important to know when your customers do something as what they actually do. You need to know how far along they are in the purchasing funnel, and what day/week/whatever along that path they’re most likely to convert.
And, of course, you also need to know which audiences to exclude. Current customers, for example, don’t need to keep seeing the same ads. Neither do people who stayed for less than 10 seconds on the site.
A note on ecommerce sites: These businesses have, at least to some extent, a more predictable customer behavior. To that end, she recommends dynamic remarketing to keep these customers interested in your products.
Finally, as part of this display marketing, Susan touched on pushing ads through Gmail. She said that these ads can work, and that while they are display-like, they’re not the same volume, so you get search-like results.
There are some different metrics to watch, too. Instead of following the same clickthroughs, you should be more interested in Gmail saves, Forwards, and clicks to the website. It may require a little more calculation to get effective data, but it’s important if this is a rout you decide to go.
While this may make these kinds of ads seem less effective and harder to use than normal display ads, it can be very beneficial if you’re targeting a mobile audience. We all check our email very regularly on our phones, so in order to be effective with these ads, we have to target people with an offer that they can easily save and use later.
In the end, it’s all about determining who, what, and where your audience is so you can be hyper focused on it.
Then test it. And test it again.
This was a great presentation to end the first session of the conference and give us something to talk about as we all adjourned for lunch. We’ll come back for the next session and see what else the presenters can bring to the table.