Our six-week experiment is over. We’ve shown you just how easy it is to start an e-commerce company and advertise to thousands of people.
We’ve also shown you our process for improving ads, where things went wrong, and what we did differently to learn from our mistakes.
But the best is yet to come!
Last week, in Part 1, we took a hard look at our processes. In this post, Part 2, we’ll give you the full rundown: the results of our latest campaign, insights from our websites, and the costs and results of everything we’ve done.
First, for those who missed the previous posts, here they are:
Our Final Campaign: The Bernie Bro T-Shirt
We started our experiment with our “L Train” t-shirt to reach a local audience. In our third campaign, we took our experiment nationwide.
The “Bernie Bro” shirt aimed at newly inspired college students caught up in the Presidential race.
So, how did we do? In our previous post, we talked about the high initial social engagement on our ad. That trend continued through our campaign. By the end of it, we had 41 likes and 2 shares.
We knew our ad was really engaging people because the comment section took on a life of its own, stoking a minor meme war between people who were for and against Bernie Sanders.
When people like or comment like this, Facebook will share that interaction with their closest peers. That leverages the network effect amongst a person’s friend group and promotes our ad for free.
This isn’t a popularity contest. But metrics suggested we were closer to the right audience than in our first two campaigns.
The “Bernie Bro” campaign cost $110 and reached 6,394 people. It had a low frequency—on average, our ad was served to each person 1.6 times. That meant we probably had a good match between our budget and target audience.
But we were still disappointed. We were getting much more traction, but it still didn’t translate into sales.
Our final campaign had one clear new lesson:
- Social engagement isn’t the same thing as hitting your sales goals. But you won’t know whether you need to choose a different audience or improve your ad design until you run more tests. A/B tests different images and language to find out whether there’s a message that will create real conversions, or whether your audience is interested but unconvinced by your product.
What We Learned from Our Analytics
We’ve focused on Facebook Ads campaigns. But to really see their impact, we have to look beyond our AdEspresso dashboard to our website itself.
There was a lot of traffic from our Facebook Ads Campaigns. All told, it drove 715 unique visitors.
The most traffic came from our first “L Train” campaign, which had huge reach through Facebook’s audience network, which broadcasts ads on apps off of Facebook. However, as we discussed in a previous post, our ads were only shown on the audience network in our first campaign. That meant lots of views but unreliable metrics, like too-high conversion numbers, that made it hard to build on.
The biggest boosts came in the first couple days after each campaign began. Looking at Google Analytics showed us the same patterns.
That shows frequent, short campaigns will drive more traffic than longer campaigns. More important, it showed we could get people to visit our site, so our audience targeting was moving in the right direction.
The Curious Case of Cart Abandonment
Shopify gave us another major insight. We only sold one shirt, but we were a lot closer to successful sales than that. 35 customers, or 4.44% of our total visitors, added one of our t-shirts to their cart.
That’s not far off from the average e-commerce add-to-cart rate of around 7 to 8 percent. But from there we had a really steep drop off:
- Only 4 people who added to their cart reached checkout (11% of total)
- Only 1 person actually completed an order (2% of total)
High cart abandonment is common in e-commerce with an average rate of about 68%. But our abandonment rate was 98%.
Why did that happen? We can’t be totally sure, but we think it has to do with hidden costs.
56% of people cite additional costs as the reason they don’t complete a purchase. Looking at our own product, that seemed plausible.
We advertised our shirts for $17.99. But after taxes and shipping, the real cost of each t-shirt was about $25. We chose a price for our shirts that put it firmly in the average for most t-shirt brands. With taxes and fees, our t-shirts had the price of a higher quality or branded product.
We could have addressed this in a couple different ways:
- Price transparency: the price on the website includes taxes and fees. That might mean fewer people would add it to their cart, but increase the number of people who complete a sale.
- Have a cart abandonment strategy. We could have used tools to win people back. Shopify has integrations with apps like Jilt and Abandon App that let you set up custom email campaigns to convert potential customers who abandon their cart before purchase.
It’s frustrating to get so close, only to end up with an abandoned cart. But we’re confident that with a few changes, we could see significant improvements in those numbers that lead to more sales.
Getting people to your website is great, but it’s just the first step. The big lesson from our Shopify analytics was:
- You need a strategy to get customers over the finish line. Our Facebook ads brought people to the website, but we lost a lot of them in the last mile. When you’re planning a campaign for your product, make sure you have a strategy in place to win people over when they get to your site via customer testimonials, free shipping, or money back guarantees, and that you have a process to reach out to people who like your product, but didn’t take the last step.
Looking Back: Our Campaign in a Nutshell
So, finally, we look back at our six-week experiment in its entirety. What did we spend? What did we achieve? What’s the state of King’s County Threads?
It’s hard to overstate the most important part of our experiment: creating a process we could use to learn from each campaign.
With our process in place, we were able to build on each campaign to improve subsequent ones.
The easiest and most cost-effective part of our experiment was setting it up. As we wrote before, we created a store, designed a t-shirt, and integrated it with Facebook for $49.87 and about an hour’s worth of effort.
Our first campaign, for the “L-Train” t-shirt, looked like this:
- Cost: $47.03
- Cost Per Click: $0.10
- Click Through Rate: 2.599%
- Reach: 7,395
- Conversions: 0
It reached a lot of people, but only on one channel, and didn’t achieve our goal. But it helped us improve our audience and narrow down our ad placements.
Our second campaign, also for the “L-Train” t-shirt, had these results:
- Cost: $88.24
- Cost Per Click: $1.36
- Click Through Rate: 0.196%
- Reach: 4,883
- Conversions: 1
Our second campaign was more expensive, and reached fewer people, but that’s in large part because it reached them on a more valuable real estate. We focused on the Facebook Newsfeed and Right-side placements, which are more expensive but more likely to reach people. It worked for one person, garnering us our first sale.
Faced with relatively high cost for each click to our website, we switched it up for our final campaign, targeting a brand new audience with our “Bernie Bro” t-shirt:
- Cost: $110.11
- Cost Per Click: $3.15
- Click Through Rate: 0.343%
- Reach: 6,394
- Conversions: 0
That meant in total our experiment cost $295.25 over its six weeks.
That would not be a desirable customer acquisition cost for any company. But even above our goal of selling t-shirts, the goal was to experiment, make mistakes, and improve on them. By obsessively documenting each step of our experiment, your business can avoid those mistakes and see better results.
No one campaign achieves huge success. Facebook ads work through iteration, learning, and revision. But each campaign came closer to finding an interested audience, and each drew hundreds of people to our website.
What Will Your E-Commerce Company Look Like?
We’ve received a lot of comments, suggestions, and general enthusiasm over the course of our experiment, and we’re excited to learn where people will take it next.
We’re going to keep King’s County Threads running—the L Train still might shut down, and Bernie Sanders is still running for President—and we want to improve our ads and website. But we’re only one company, with one kind of product. The barriers to entry for any e-commerce company are incredibly low, and there are so many opportunities to find new audiences on Facebook.
We’ve shared what we did. Now, we want to know what processes you might create for your own company, and what sort of ways you can build on our experience to drive your own success.
Add your comments and let us know your plan for e-commerce success!