Blog: Observations & Answers

It's been an amazing 24 hours for the Campaign Monitor team since we launched fixoutlook.org on behalf of the Email Standards Project yesterday afternoon (Sydney time). In less than a day, we hit an incredible 20,000 tweets from the community who shared our position on Microsoft's lack of standards support in Outlook 2010.

Just as impressively, we received an official response from Microsoft before the day was out. While this wasn't exactly the message we'd been hoping for, there were certainly some positives to be taken from it. We posted a follow up to Microsoft's response on the Email Standards project blog not long after.

Twitter is by no means a vacuum, and the amount of tweets and interest in our approach resulted in coverage across the web on a huge scale. Here are some of the highlights.

• Microsoft, Outlook Is Broken, Says 6,000 Tweets (And Growing). Fix It. - Techcrunch
• Users Pressure Microsoft to Fix Outlook - Reuters
• Microsoft defends Outlook HTML decision - CNET News
• Microsoft: Outlook's not broken and we aren't 'fixing' it - ZDNet.com
• Outlook 2010 not winning fans. . .on Twitter - CNN.com
• Sour Outlook - zeldman.com

This coverage combined with the relentless flow of tweets resulted in more than 100,000 page views on fixoutlook.org in the first day.

The story behind the site

The moment we heard the news about Outlook 2010 and tested it for ourselves, we knew it was important to act fast. If there was any possibility of getting Microsoft on board before Office 2010 then we couldn't delay any longer. After a brainstorming session, the idea to use Twitter to spread the word was born and developed quickly from there.

Ever since the site launched, we've had loads of enquiries from people interested in how the site was actually built. The site was created in less than a week by the talented guys at Newism, which we've worked with on loads of projects before. I think we can all agree that they did an amazing job.

It was a completely home grown solution the guys built out of the box using the Twitter search API. The Newism team have told me they plan on doing a more comprehensive post on the nuts and bolts of this process, so I'll be sure to share that when it eventuates.

While Twitter has been used to encourage change in lots of different ways before, there were a few unique additions that I think contributed to our success. Todd Zeigler summed it up perfectly in his post today.

What's next

We've already been back in touch with Microsoft since the site launched, and are hoping that the continued pressure from the community will lead them to re-assess their position on standards. In the mean time, we'll keep trying our best to spread the word through the ESP blog and the newly created @fixoutlook account on Twitter.

Thanks again for the amazing level of support we've had to date. 20,000 people coming together to encourage change in 24 hours is something we can all be proud of.

We'll be performing a database upgrade this Sunday which will result in a 3 hour downtime period for Campaign Monitor this Sunday evening (June 14) at 5pm US Eastern Daylight time (see this in your own time zone). This upgrade is part of a sweeping round of speed improvements we have planned for Campaign Monitor over the coming weeks.

What will this mean for me?

This maintenance period will have the following effects:

• You and your clients will not be able to login to your Campaign Monitor account. If your client does try and log in they'll see a non-branded "Down for maintenance" page letting them know when they should come back and try again.
• Campaigns scheduled to be sent during this period will not go out, but will be sent automatically as soon as the application comes back up. However, subscribes, unsubscribes and bounces will be captured during the downtime and processed into your reports afterwards. Images will continue to load in emails, and clicks and opens will also be recorded.

We will update this post once the application is back up and running and we'll also be updating our Twitter status throughout the maintenance to keep you in the loop. Thanks in advance for your patience while we complete this upgrade.

Update: The maintenance has been completed and was a success, and we have re-enabled account access. Thanks for your patience everyone.

Back in April of this year we decided to have a look at running some banner ads on a number of high profile design-related sites. Over the last couple of years, advertising in our little corner of the web has come a long way. There are loads of highly trafficked design sites with excellent content and affordable ad slots. On top of this, the rise of targeted ad networks such as The Deck and Fusion Ads has made it much easier to get in front of the right crowd.

While we've dabbled in some banner advertising before, I decided to take a more thoughtful approach this time. We put together a number of banners, dedicated landing pages and put conversion tracking in place to measure the results. When looking into this process initially, I didn't come across many write-ups from advertisers on what worked, what didn't and just how effective the ads had been for them. In the interest of helping fill that void, here's the process we went through, and some of the surprising results that eventuated.

Coming up with the ad creative

As part of the merger between Campaign Monitor and MailBuild last year (more on that here), we added a stack of new features to make it easy for designers to earn passive income off their clients through email marketing. In a nutshell, you can create a sub-account for each client, set the price they should pay and earn a profit every time they send.

We figured this was something that lots of designers might find useful and it became the focus of our banner ads and associated landing pages. Here are the 3 ads we ran with initially.

For the final design of these ads we enlisted the help of the clever team at Newism, the same team that coded the current version of the Campaign Monitor site. We couldn't have been happier with the results.

The Landing Pages

As well as testing multiple ad creative, we decided to put together a number of dedicated landing pages for the campaign. We'd randomly display a different landing page for each visitor and measure which one converted best. We bought in the amazing Jesse Bennett-Chamberlain (who designed our web site) and Brad Hayes to help with the design.

Landing Page 1: Earn money in your sleep

This key for this page was to promote the idea of earning money in your sleep. We highlighted the 3 simple steps involved (rebrand, resell and then profit) and included an inline signup form allowing those interested to signup and start using the app on the spot.

Landing Page 2: Are you a designer?

This was definitely the most controversial landing page of the four. We took the fact that Campaign Monitor is built for designers to the extreme by including a modal window that overlays the landing page asking the person if they are a designer or not. If they clicked "You betcha", they'd be shown the page below that highlights how easy it is to resell Campaign Monitor.

If they clicked "Nope", the landing page behind the modal would fade out with a message explaining that "It's not you. It's us." We then linked to a number of other Email Service Providers that they might consider checking out that are built for a less targeted audience.

Unexpectedly, this approach stirred up a lot of conversation on Twitter. Some called us ballsy and classy for linking to our competition, others asked if we were breaking up with them. I personally received a number of emails from people complimenting us on being open and honest with people. The truth is, we've always referred potential customers to our competitors when we know they're not the right match (see this post from 2 years ago). It's how we'd want to be treated, so it's only fair that we do the same to our customers.

Landing page 3: Modal-free just for designers

In the interest of testing what sort of impact the modal window might have on conversions, we also tested it against the very same landing page modal-free. The page was exactly the same otherwise, so should provide a nice comparison.

Landing page 4: Our home page

In the past we've just pointed people to our home page, which in itself is a landing page aimed at converting people to give us a try. Will it out-perform the dedicated landing pages that have a clearer connection to the ad creative?

The Results

Over April and May we booked advertising spots with The Deck and Fusion Ads along with Smashing Magazine, Six Revisions, Webdesigner Depot, Webresources Depot and CSS Elite.

We set up a redirect script on our servers that would choose one of the 4 landing pages above for each visitor and pass through any of the required parameters in the URL so we could track everything with Google Analytics. We have a number of goals set up so we can track important things like a customer signing up, sending a test campaign and becoming a paying customer. If you're interested, our designer Dave Martin has written about our Google Analytics setup in more detail here.

We judged the performance of each banner ad and landing page on the number of visitors that signed up to Campaign Monitor, as opposed to basing it purely on revenue. Because of the nature of our pricing, it can be weeks or even months before a customers starts paying for our software. We'll use the revenue numbers internally over the coming months to get a true idea of ROI.

The best performing banner ad

The blueprint banner ad outperformed the other two with a conversion rate of 3.5% resulting in 370 people signing up for Campaign Monitor. Here are the full results.

Banner Ad	Conversion Rate
Blueprint	3.51%
Post-it Note	2.36%
Chalkboard	2.09%

The best performing landing page

This one surprised everyone. Personally, I was concerned the modal window would result in a lower conversion rate because of the barrier of an additional click. I was pleasantly surprised to see this page gave us 25% more conversions than the next best performing page. Possibly the biggest surprise of all was that the "Earn money in your sleep" page with the inline signup form didn't give us one new customer. Not one! Here are the full results.

Landing Page	Conversion Rate
Landing Page 2: Are you a designer?	4.34%
Landing page 3: Modal-free just for designers	3.48%
Landing page 4: Our home page	1.52%
Landing Page 1: Earn money in your sleep	0%

The best explanation we could provide for this is that people like to check out a product more before signing up (there was no link to the product from the page, the focus was on the signup form). The other landing pages provided links back to the site where an interested customer could take a feature tour, check out the pricing, etc before signing up. On top of this, the form had a total of eight required fields. Reducing this to the bare minimum and asking for the customer data after they signed up might have helped convert more. But still, not one?

The best performing advertiser

From the day we turned these ads on, Fusion Ads has consistently been our best converter. This includes visitors who have signed up right through to total revenue to date. We've also seen good conversion rates from other advertisers, which you can see in the results below.

Advertiser	Conversion Rate
Fusion Ads	3.17%
The Deck	2.62%
Smashing Magazine	2.46%
Six Revisions	2.05%
CSS Elite	2.03%
Webdesigner Depot	2.02%
Webresources Depot	1.9%

It's important to keep in mind that these are conversion rates only, and don't give any indication of the true return on investment. For example, while The Deck has been our second best converter, it's also more expensive than any of the other ad slots. Because of the rates we've managed to negotiate with some providers, I'm afraid I can't share the costs associated with each ad spot, but most are easily found on each respective advertiser's site.

Was it worth it?

Right now it's too early to tell just how long it will take for this campaign to pay for itself and then turn a profit (don't forget the cost of designing the banners and landing pages as well as the ongoing advertising fees). To date we've only recovered around 25% of all expenses in revenue from new customers. Having said that, a large percentage of our customers continue to use our software for an extended period of time. Looking at how revenue has been growing from these customers in the last few weeks, it certainly seems that in the longer-term this exercise will be well worth it. The nature of our business model means that it will be a few more months before we'll really know.

To me, the most important element was actually testing everything as we went. By putting in a little extra work, we could quickly gauge which ad creative, landing page and advertiser was giving us the best results and react accordingly.

Our next steps will be to continue to refine the winning landing page using Google Website Optimizer to improve conversions further. On top of this we'll be trying some new ad creative and throwing a few new advertisers into the mix. I'll try and put together a follow up in a few months time with anything else we might have learned in the mean time.

After a huge 122 suggestions, the time has come to pick some winners for the latest Campaign Monitor t-shirts. We've got together and picked out some finalists for you all to vote on, so now it is your turn.

What we were looking for was something fun, something related to Campaign Monitor or email marketing, but also something that wasn't just a promotional shirt. There is enough of those around already.

We wanted something people would actually enjoy wearing, not just a billboard for your chest. It should also be something that lends itself to a t-shirt design, and isn't too insulting.

We've also added in some of the popular suggestions that didn't quite win last time, to give them another chance at glory.

We'll leave the voting open for a week or so, and announce the winning shirts some time after that.

Thanks to everyone who contributed suggestions. Now go add your vote!

Yesterday morning, Sydney time, our IP addresses were added by a major blacklist provider to one of their lists. We discussed the problem in more detail in our blog post to keep you all up to date. We also posted updates and answered questions on our twitter account.

We just wanted to give some more details, most importantly to let you know that the block was removed.

We are no longer blacklisted

We finally heard back from Spamhaus late last night, letting us know they had removed the listing. At that point, many of the blocking problems were instantly resolved, including Yahoo mail. However, a lot of smaller ISPs and email providers maintain their own lists, and use Spamhaus as one of their sources.

So our support team then ensured that we contacted and were delisted by all of those companies too. A special thanks to Diana, support team legend, who did a lot of that work.

The end result is that the vast majority of blocks have been resolved now. We are still waiting for some of those smaller delistings to trickle through and take effect, and expect that to happen sometime today. If your campaign is not urgent, it would be best to wait a little longer to give those providers more chance to update their lists.

Please contact us for refunds of the blocked emails

If your campaign (or your client's campaign) was impacted, please contact support and let us know. We will add credits to your account to cover the actual blocks, and also the sending fee, so you can resend the campaign to the affected subscribers at no cost.

How can I resend to the people who were blocked?

Unfortunately, we don't have a direct way to do this for you. You have a couple of options, depending on the size of the campaign and which domains were blocked for you.

Create a segment of the blocked domains

If you look at your campaign report, and you can see that you had soft bounce blocks from up to (say) 10 different domains, you can send just to those people. To do that, first make a list of the domains you want to send to.

Then, jump into the 'segments' section of your list. See our help topic on segments for an overview of the process. For this method, you'll want to create a segment based on email addresses.

Setup a series of OR rules like

Email address CONTAINS yahoo

OR

Email address CONTAINS hotmail

Adding in each of the domains you want to resend to. This method will be the most accurate way to send to the people you know did not receive it. If you have too many domains to deal with, you'll need to use the second method:

Create a segment of all unopened emails

This method is much faster, if you have people from a lot of different domains being blocked. Again, you would create a segment, but this time using a rule like:

Campaign (your campaign name) WAS NOT opened

If you use this method, it is quite likely you will resend to some people who did actually get your email (see how open tracking works), so you may want to add a notice to the top of your email apologising and explaining the situation.

Exporting the bounces and creating a new list

This technique is quite simple and effective as a one off resend. You can view your affected campaign, and click through to the soft bounces. If you see that most or all look like blocks, you can export the list of bounces from that page. Then create a new subscriber list, and import the file you just created. That gives you a list you can send the campaign to, and then delete afterwards.

The choice of which technique to use really depends on the size of your campaign, how heavily it was impacted and how urgent the material is. Keep in mind that you may not need to resend at all. It is unlikely that every person who was blocked would have actually opened the email, so you may decide just to send your next campaign to everyone as per normal.

These blocks were only soft bounces, so the subscribers will still be active in your list. We'd like to apologise again for the trouble this has caused, and please know that we are already working on ways to avoid being in this position again.

We'd also like to thank everyone who was supportive and understanding during this period, all of us at Campaign Monitor really appreciate it.