What is Remarketing?

First let me explain what remarketing is (or my understanding of what it is). Remarketing is a cookie-based system that lets you show your ads to people who have visited your site, but have not purchased yet. Your ads can appear on any site that the user visits, provided that your ad company has an arrangement with that site. And that’s the beauty of adwords now offering remarketing: its content network is huge.

Here’s an example:
A user visits my Comic Collector product page. Maybe he clicked my search ad on Google, or an image ad on the Content Network, or the Comic Collector listing in the organic results. It doesn’t matter really, he may even have come from Bing, or a link on a comic forum discussing cataloging software. This user doesn’t buy and leaves the site (annoyingly, that still happens to us sometimes ). Maybe he downloaded the trial edition, maybe not.
Now after he visits my website, this user goes on to browse the internet, visiting other websites. And suddenly everywhere he goes, he sees Comic Collector ads. Site-wide leaderboards, towers, inline rectangles, even little text ads. My ads are following him around, almost stalking him.

One important thing to understand: remarketing ads do not only appear on websites related to your product, like your “normal” content network ads. No, they appear everywhere, on any site.
Remarketing let’s you target specific users wherever they go, as opposed to all users on specific websites.

How does Remarketing work?

As I said, it is a cookie-based system. Every user that visits your website (or a specific section of it) gets cookied by your ad system. Then, when the user leaves your site without buying, this cookie is read by the same ad system on other websites. And if your bid is high enough, the ad system then shows your ads to that specific visitor. As soon as your stalking makes him give up and finally buy your stuff, he is un-cookied and he will not see your remarketing ads again.

Quick guide for setting it up in Adwords

Remarketing is not something you just “switch on” in your Adwords account. There’s two main steps to complete: Creating lists and then targeting those lists with ads.

Creating a Remarketing List

• Click the “Control Panel & Library” link at the bottom of your Campaigns list on the left, and click Audiences.
• Click the New Audience button and choose Remarketing List.
• Give it a name and choose a “membership” duration. (It looks like Adwords does not do the un-cookying bit, it simply stops showing your ads after a predefined number of days)
• Save your remarketing list. You will see it appear in the list.
• Click the link in the Tags / Rules column, to get your Remarketing code ( a piece of tracking Javascript).
• Paste this piece of code right before the ending “body” tag of the pages whose visitors you want added to this particular list.
• That’s it. All visitors to those pages will now be added to your Remarking List. You can see the number of users that are on the list on the Audiences page.

Targeting your Remarketing Lists with ads

Targeting a list is done using the new Audiences tab that is now present on any ad group page. You can add your list to any existing ad group, which will show that group’s ads to your defined remarketing audience. But I’d recommend to create a new ad group that targets only the remarketing list, so no keywords, no placements. This makes it easier to track the results.
And what’s more important: you probably want to show different ads here anyway (see “How to use Remarketing” below).

• In your ad group, go to the Audiences tab.
• Click the Add audiences button.
• Now select the Remarketing List you created above and click “add”, then Save.
• Then go to the Ads tab and create some ads for the new audience.

How to use Remarketing?

Well, I am still trying to get my head around that, but here’s some ideas.
You can of course simply show your regular ads to the remarketing audience. That should be somewhat effective, as it can remind those users of your product or website.

But as you are targeting people who have already visited your site at least once, you can assume that they already know what your product is about. So there’s probably less need to explain your product to them.
You can use an entirely different message now, focussed on getting them back to your site. You can go for a soft reminder, showing a branding-type of ad. Or you can be more direct, offering them a special discount if they buy now, making the ad click directly to your check-out page.

Currently, I am hesitant to make it too obvious to users that they are remarketed to, that they are seeing your ads because they visited your website earlier. Privacy-sensitive people may not like that. I mean, you could go as far as showing them an ad that says “Hey, remember Comic Collector? Please come back, and I’ll give you an extra discount if you buy now”. Tricky?

This is what I am trying now for Comic Collector:

For comparison, this is one of my regular image ads for Comic Collector:

Results?

Well, I started two days ago, and my new image ads have not even been approved yet. But I am seeing my Remarketing Lists filling up with users already (around a thousand each now).
My guess is that remarketing will become more and more effective as your lists grow. I have set the membership duration of my lists to 60 days, so one would expect the lists to stop growing after 60 days. You could set the duration even longer to let the target audience grow bigger, but I wonder how useful that would be. For our products, if people haven’t purchased after 60 days, they won’t.

I will let you know as soon as I am seeing results, as in actual sales. The special coupon should make it easy to spot “remarketed” sales.

“Sell benefits, not features”

In general, this is good advice. But in the software world, it is not always the best approach. In some situations, pushing your features will sell better. I encountered a situation like that a while ago.

It all depends on your product, your target audience and your sources of website traffic.

Product

Let’s say you are launching a completely new type of product, a smart piece of software that solves a problem that is not being solved by any other program. Then yes, you will need to describe the software in terms of benefits. Explaining which problem it solves, describing the benefits of using the software and why it is so great to have that particular problem solved.

But for most of us, the software we sell is not that unique. You probably have several competitors who make the same type of program. Examples: Text editors, FTP clients, file managers, image editors, etc…
You may even be up against the big guys, creating your own email client, MP3 player or file compression software.

In that case, there is less need to describe the benefits of using the software (if at all), especially if you can assume that most of your visitors are aware of the competing products (which of course also depends on your audience and your traffic sources).
If your average visitor knows about the other programs, you want to push the features that set you apart (because of course your software is better). If you are competing against one or more well-known competitors in your market, you can go as far as creating a feature comparison table between your program and those alternatives.

Target Audience

If your audience is tech-savvy, then selling in terms of features can work better too. These visitors don’t need to be told about the benefits, they already know them or can easily deduct them from the features list.

And then there’s the ultimate tech-savvy audience: developers.
Are you creating developer tools? Then an honest and to-the-point feature list may be the best way to sell your software. These guys may even be allergic to benefit-style bullets

Computer-savvy / internet-savvy users are also more likely to be aware of alternatives and your competitors. They have probably done several Google searches already, found several options and are now visiting your site to determine how well you compare.

Traffic Sources

Which automatically brings us to the third factor: the sources of your traffic. Where do your visitors come from?

If they arrive at your website after a laser focused Google search, clicking on your Google ad, then you can safely assume that they know exactly what they want and why they want it. They don’t need to have the benefits explained to them. And Google has probably given them a nice list of your competitors too.
So for that type of visitors, focus on explaining why your features make your product the right tool for the job.

On the other hand, if you have a lot of traffic that is less targeted (e.g. from Google’s Content Network, from banner ads, organic search engine listings) then you may have some benefit-explaining to do.

For example, let’s say an ad for my DVD cataloging software shows up on IMDb through the Content Network. IMDb’s visitors are interested in movies, so they are somewhat targeted. But they may have never considered cataloging their DVDs on their computer.
So if those visitors click my ad and arrive on my Movie Collector product page, I need to explain to them why they would need (or at least want) to create a personal DVD database. E.g. to prevent duplicate purchases, to make sure they don’t loose sight of loaned DVDs.
And I can’t suffice with telling them that my software automatically downloads all DVD details and cover images (which is the main feature). Instead I will need to push the benefit of being able to catalog your entire DVD collection in just a few hours, without typing.

Now the above two traffic sources are two extremes. In practice, it isn’t this clear-cut.

Looking at keyword-driven search engine traffic, visitor types are usually different depending on the keyword.
Some keywords bring highly targeted visitors that need feature lists. Other keywords brings visitors that are still in an early phase of their search and you may draw them in by telling them about the benefits of your solution.

Let’s take a look at a Movie Collector example again. Consider two searchers, one searching Google for “dvd database software”, the other typing “movie collection” into the search box. I use Adwords to bid on both keywords, with separate ads of course. Now these visitors are two completely different kinds of animals.

If someone searches for “dvd database software” and hits my site, do I need to bother him with the benefits of having his DVDs cataloged in a database? Of course not, he knows that he wants and why he wants it. He just needs to know that my product is the best software to do the job. A feature list will do just that.

A “movie collection” visitor is different. He probably owns a lot of movies, collects them even, but he has not yet realized that cataloging his collection would solve his problem. In fact, he may not be aware that he has a “problem” yet. So let’s tell him about his problems first, then present the solution with all its benefits

Sell features or sell benefits?

Now what should we do, sell features or sell benefits?

Easy: just do both.

Create two landing pages for each product. One explaining the benefits of your solution, one clearly listing its features. Then send your traffic to one or the other based on the traffic source and the keyword.

Not all traffic can be funneled to the ideal page, but your Adwords traffic can.
That is, if you have your Content Network and Search Network nicely separated into different campaigns and if you have your keywords grouped into focused ad groups with their own landing pages. But of course you have…

Version A, emotional, benefit-oriented:

Starting right now, you will enjoy a perfectly organized DVD collection…

Movie Collector Instantly Catalogs Your DVD & Blu-Ray discs.

You’ll Save Time and Money Today.

Finally Have Total Control Over Your Movie Collection.

Version B, practical, feature-oriented:

Automatic DVD Database Software for Windows, Mac OS X & iPhone

Movie Collector Automatically Catalogs Your DVD Collection.

Just Enter Movie Titles or Scan DVD Barcodes.

Instantly Download All Movie Details and Cover Images.

Seven days of testing gave me the following interesting results:

• Downloads: up 1.5%
• Sign-ups: up 2.3%
• Sales: up 9.7%
• Average First Purchase Value: up 13.5%
• Profits: up 24.4%

I was expecting version B to do better. People who are into cataloging their stuff tend to be a serious, practical crowd. But I wasn’t expecting the difference to be this big.
I am not complaining though

Now I must say that most of my current traffic is search engine traffic, either organic or paid search, e.g. searching for “dvd database software”. I suspect that this type of visitor knows exactly what he wants to do and why. He or she is just searching for the best software to do it. They don’t need to be conviced that they should organize their stuff to save time and money and to get control over their collection. They just want confirmation that Movie Collector will do what they want as efficiently as possible.

For other less targeted traffic like Content Network traffic, results may be different.
So I unpaused my Movie Collector Content Network campaign and am testing that now. Who knows, with a special fluffy landing page the Content Network may finally become profitable.

I have been using the new UI for a week now. Here’s my first impressions:

First, here a couple of things I really like about the new UI:

• Graphs on each page: Makes it a lot easier to spot trends in clicks, impressions, CTRs, etc…
• Navigation bar on the left: Quicker navigation between campaigns and ad groups
• In place editing: edit bids, keywords, ad group names etc… without going to a different page (and back)
• New “See Search Terms” feature: This gives you Search Query information straight from the Keywords tab. Great for finding negative keywords.
• New “Networks” Tab shows “Automatic placements” for the Content Network. Remember my post about Keyword Targeting vs Placement Targeting for the Content Network? I mentioned that one of the annoyances of Keyword Targeting was that you had to run Placement Reports to find out where your ads are actually being shown. Not anymore. You can now find this information directly within your ad group, on the Networks tab. You can select sites right there and choose “Exclude”. Great!

Of course, I also had some annoyances:

• Missing Next/Previous links: In the “old” UI I always use the the little Next/Previous links to navigate from one ad group to the next. The nice thing is that these follow the order you had your ads in at the Campaign page. These links are gone now. True, you can now use the nav bar on the left, but strangely enough the ad groups are always ordered by name there…
• Less usable on lo-res screens: I am doing most of my Adwords work on my little MacBook which has a 1280 x 800 screen resolution. The new UI causes more scrolling, even some horizontal scrolling here and there. It’s a little better when I had the nav panel on the left, but still not ideal.
• Missing Edit Keywords box: The new UI doesn’t seem to have the Edit Keywords box that lets you edit your list directly in a simple text editor box. I am always using this to copy / paste my entire list into external keyword tools and back. I am really going to miss that one.
• Settings not remembered: In several places, the UI doesn’t seem to remember my settings, like sort orders, graph settings, etc… I am sure this will be fixed before the public launch though.

At the moment, I find myself switching back to the old UI to do particular tasks. But as soon as the new UI goes live, I won’t be able to. I guess we will just have to hope that the above annoyances will be fixed before then.

Up till now I have been using keyword targeting only. And I have been running Placement Reports to find “bad” sites, so that I could exclude them using site exclusions.

But now my campaigns have run for a few months, my ads are being shown on thousands of sites. There is no way I can go through and exclude the all bad ones.
I have removed the high volume bad sites which helps a bit, but there’s load’s of bad ones in the long tail too. Some of em even generating lots of clicks and thus costing me money.

(BTW: when I say “bad sites”, I mean “somewhat relevant” sites that are not performing plus sites that are simply not relevant.)

So now I am trying placement targeting. I am using the Placement Report for my keyword targeted ads to find good sites (based on relevance, number of conversions, clicks and impressions). Then I create new placement targeted ad groups to target just those sites.

Here’s my findings on both methods:

Keyword Targeting

• Easy to setup: just create an ad group with a tightly focused set of keywords. Then let Google find relevant sites to show your ad on.
• However, Google is likely to show your ads on less relevant sites too. Which may severely affect the effectiveness of your campaign.
• So you need to use Placement Reports to find the bad sites and exclude them using site exclusions. If your ads are shown on many sites this can be a lot of work. Too much work in my case. Maybe my keyword groups are not focused enough?

Placement Targeting

• Takes a little more work to get started, as you have to find relevant sites yourself.
• Having run keyword targeted ads first can be very helpful, so that you can use your Placement Reports to find good sites to target.
• But I find that the Placement Tool does quite a good job too. For me, the best way to get started was to use the List URLs option and then just plug in a couple of relevant sites. The tool then automatically gives you related sites.
• This process will automatically make your advertising more targeted and probably more cost effective.
• However, you may miss the long tail of many small relevant sites.
• In general, the average cost per click seems to be higher when using placement targeting, even on the same sites.
• Some sites are “not available for placement targeting”. Even while they showed my ads when I used keyword targeting, i.e. gamespot.com. Is that an option in AdSense?
• One big advantage of placement targeting: you don’t have to use reports to get the basic performance data (impressions, clicks, conversions, etc..). This data is immediately available on your ad group page, on the Placements tab.

Adwords also lets you combine the two, still experimenting with that.

Preliminary results: impressions are way down, click through rates are up, conversion rates are up, costs are down.

Of course, when I started out with Adwords in 2002 I made the same mistake everyone makes: leaving the Content Network switched ON for all campaigns. Which got me lots of impressions of course, many clicks even, some (expensive) sales too.

But reading up on Adwords I learned that one should always have separate campaigns for Search and Content Network.

So I did exactly that. I switched off the Content Network for my campaigns and created separate campaigns targeted at just the Content Network. I created some ad groups with my main keywords (the same ones I was using on the search network).
Click through rates for my regular campaigns went way up, so that’s a good thing. But on the content targeted campaigns I got impressions (loads of em), costly clicks too, but almost no sales at all.

So I paused the campaigns and just gave up on the Content Network.

A few years later, when Google introduced site-targeting, I figured I’d try my luck again. So I manually selected some sites I thought would be targeted for our Movie Collector software. A couple of DVD review sites, online DVD shops, home theater sites, etc… This time results were a little better.
I got some sales, but the cost per conversion was way too high, mainly because I had to bid high to even show up on those sites. Conversion rates were terrible, up to the point of paying $100 to get one $40 sale. Not what I had in mind, so I gave up again.

However, a few weeks ago I listened to Perry Marshall’s “Adwords Expert Series”, a series of interviews with Adwords experts. One of these experts was Shelly Ellis, an expert on Content Targeting.

And Perry’s interview with Shelly contained some interesting eye-openers. Here’s what I learned:

• Ad groups for the Content Network should have a tightly focused theme, using only a couple of keywords (5 to 10).
• The keywords should be used to target the theme of the sites you want to be shown on. Often these are completely different from your search network keywords.
• Make sure your ad matches the audience of the sites and your landing pages matches the ad and the audience.
• You can use Placement Reports to see where your ads are showing.
• Use Site Exclusions to remove sites that have low click through rates, or worse, high click-through rates but no conversions.
• It is possible to use negative keywords to prevent being shown on “bad” sites.
• Often, the content network may result in “indirect” sales. E.g. people see your ad, maybe multiple times, and then visit your site later by searching Google on your brand name. (which may make it difficult to track results)

So I’m at it again. I have created new campaigns, one for each of our programs, each having several tight ad groups with only a few keywords.
For Movie Collector I am trying to target DVD / Blu-Ray review sites, using just “dvd, dvds, blu-ray, bluray, etc…” as keywords.
For Game Collector I am also targeting review sites, with even tighter ad groups, one per game console (wii, ps3, xbox 360, etc…).

And I am already exclusing lots of sites that are not relevant or just plain bad sites (sites offering cracked software for instance).
The problem is that for Game Collector and Comic Collector, my ads are showing up on over one thousand sites. There’s no way I can manually check em all. I will have to include or exclude them based on performance (clicks, conversion) and it may take a while before I have enough data to do that. In the meantime, I may be loosing money, but hey, information is expensive…

Or use negative keywords to try and exclude them. So far, no success with that yet. For instance, I added “themes”, “hacks” and “wallpapers” as negative keywords to my Game Collector Content campaign and my ads are still showing up on useless “PSP theme” sites.

So far, I am not making money on the content network yet.
Results are promising though. Getting millions of impressions, thousand of clicks and some sign-ups even. But for some reason, the sign-ups don’t convert to sales. At all… Which I find strange. People are interested enough to sign up for our free trial but then none of em buy (FYI: for our search network ads 25% of sign ups end up buying).

But this time, I am not giving up. I will continue to tweak and tune.

I will keep you posted about my progress.