Good and Bad at Shopping.com

We think Shopping.com is a great place to advertise and we've been drawing traffic from there for a couple of years.

The Shopping.com website is pretty much what you would expect from a shopping site. It looks a lot like Amazon or eBay with categorised pages of product listings and various search methods. The difference is that no transactions take place at the Shopping.com website. Listings link to an external website page specified by the advertiser (ideally the page where you can buy the product) and the advertiser is charged per click.

Initially it can be quite laborious to construct a datafeed of product details, but once that is mastered it is very simple to run a continuous campaign. We've been bidding the minimum $0.20 per click in the clothing category and receiving a steady stream of traffic.

Each product is separately listed with a searchable title and description. Each has a product photo, price, reviews and information about the seller's reputation. This means that a shopper has a lot of information about the product before clicking.

Shopping.com listings rank well in the search engines and are distributed to a number of affiliated shopping sites. If one wanted to get fast search engine ranking for a new site, it would be a great place to advertise.

I must confess that I have only just installed the code snippet which allows us to track sales conversions from their traffic. I've also installed the code that invites buyers to review GMtee (Shopping.com sends the buyer an e-mail a specified number of days after the transaction) and that will allow us to build a reputation. Now that this code is in place, I'm increasing our bids to a maximum of US$0.25. I'll let you know results in a month or two.

Now for two words of criticism: customer service! When the new GMtee website went live, I had to upload a revised datafeed. There was a problem processing the feed which resulted in many listings displaying half a photo, or no photo at all. Also they were still listing products that had been removed from the new datafeed.

I sent them an e-mail asking for help. Their performance pledge is to reply within seven working days, which seems about five days too many to me! We eventually received a form reply advising us to check our datafeed. I wrote back to assure them that our feed had no errors. Another week-and-a-half passed without reply.

I sent a third desperate plea for assistance and received a reply this morning. They say they've fixed the problem and that everything should be OK sometime in the next couple of days. Even if the problem is solved now, it's taken a month to communicate and that's just too long.

About Pay Per Click PPC Advertising

Following the redesign of the GMtee website and the release of our new collection, my focus has been on increasing traffic.

Traffic is not free and the bottom line is that a successful website must make a profit greater than the cost of it's traffic. In the jargon it's called "return on investment" or ROI. Ultimately internet marketing is all about generating "positive ROI" and getting the biggest bang for your buck.

If you spend $100 on traffic and make $50 in profit, you're on the road to ruin. A large company with deep pockets might sustain a negative ROI as a strategy to acquire long term repeat customers, but I would prefer to at least break even!

However, if you spend $100 on traffic and make a profit of $150, truly you have attained marketing nirvana. You can quickly increase your spending on advertising and watch your business blossom. It's not difficult to get traffic to a website, if you can afford it.

The most immediate and direct source of traffic is pay-per-click advertising. For example, GMtee advertises with Google Adwords. A Google search returns both natural search results (in the main body of the page) and sponsored search results (in a column on the right and sometimes also at the top of the page). With our Adwords account we bid against other advertisers to have our ads shown on search results for our specified keywords.

When somebody clicks on our ad they are taken to our home page and we pay a fee to Adwords. The cost per click varies for different keywords and we currently pay an average cost per click of US$0.24. If we bid higher for the most popular keywords our traffic would increase and so would our average cost per click.

It would be easy to write an ad to generate traffic. If GMtee advertised "Everything free: Today only!" and paid for top placement, I'm confident we'd have plenty of visitors. But the object of the exercise is not just to generate traffic, it is to generate qualified traffic. GMtee shirts are high-end so it would be wasteful to attract very price-conscious consumers. A good PPC ad accurately describes exactly what is on offer.

Our Most Popular Design

There's not enough pictures in this blog! Man cannot live on text alone.

This is our best-selling T-shirt (US$44.95 + $5.00 shipping) : Thunder Dragon long sleeve tee, crew neck.

A shirt like this is very expensive to manufacture. Because it's a very big print we must pay for very big films and very big screens. Thunder Dragon is printed in only one colour though, so that helps keep these costs down. Some GMtee shirts have five or six colours.

If we were manufacturing large numbers of shirts the cost of films and screens would be less significant per unit, but GMtee orders only a limited edition (usually 120) of each new design.

The most costly shirt in our new collection is the Japanese Castle shirt with long sleeves and henley collar. You wouldn't believe the manufacturing cost even if I told you. US$13.50.

More About Stumbleupon

A couple of weeks ago I started advertising GMtee at Stumbleupon. Here's my previous post about testing the appeal of our new site.

So far we've spent US$400. At $0.05 per click that's 8,000 unique visitors. Unfortunately we can't track conversions from Stumbleupon and I've seen no evidence that any of those visitors made a purchase. While it's possible that one or two sales resulted, I am sure that we came nowhere near recovering the cost of the traffic from direct sales.

Their campaign management interface is simple to use, but reports too little information about results. I had set up ten different campaigns with variations on age, gender and country. That is, GMtee home page was shown to either men or women in one of four countries: USA, Canada, UK or Australia.

Some campaigns were restricted to 18 - 55 year old Stumblers. These produced less traffic because Stumbleupon doesn't always know the age of it's members. Most traffic came from the US with only a small contribution from the other three countries.

I was disappointed that the ratio of "thumbs up" to "thumbs down" was not reported individually for each campaign. Instead the report showed only one overall approval rating (which held steady around 75%). I could have calculated individual campaign ratings from other information provided, but frankly it didn't seem to be worth the effort.

I had already concluded that our return on investment at Stumbleupon is unlikely ever to be positive. As a bonus though, I think that GMtee will continue to receive traffic even after we stop advertising, since it is in their own interest for Stumbleupon to include highly rated sites in their rotation. From time to time I will masquerade as an impartial stumbler and give the thumbs up to our own pages!

There's an intangible benefit from promoting GMtee to thousands of stumblers. Hopefully we are noticed by people who are also active on other social sites and it's possible that we could benefit from some viral networking. (That assumes you think GMtee is cool enough to tell your friends!)

Something odd happened. During the first few days of the campaign, GMtee was receiving between one and two thousand visits daily. Then the traffic dried up even though our budget had not been spent. I tried creating a new campaign in a different category (clothing instead of fashion), but it was not approved for some reason.

Next I created a new campaign in both categories, but this time pointing to our Men's T-shirts page. I restricted the traffic to men in the US. Almost 1,000 visits have been recorded and the approval rating by this group for this page is up to 85%. It's not a big sample, but woohoo! So now I've deposited another $100.

Webhost Rant

Some of you have been complaining that you can't access our site. I apologise for that.

We've been having problems with our webhost Jump Web Services They in turn are blaming the problems on the data centre.

Last night, around 1 am, I noticed that GMtee was down again. I fired off a terse e-mail to customer support and was pleasantly surprised that the problem was quickly fixed despite the ungodly hour. But when I returned to my computer in the morning the site was unavailable again.

Selling you a shirt ain't as simple as it might seem. Design, manufacture, delivery, photography, website updates, marketing... all that effort and finally our webhost can't keep their server working.

Imagine my frustration! Customers complaining. Lost sales. Damage to our reputation. Money spent advertising a dead site. Negative impact on our search ranking. Not cool.

It's expensive and inconvenient to move the site to another webhost, but it looks like we may have no choice.

/End Rant.