It recently was revealed that mobile ads account for 41 percent of Facebook’s revenue. That came as no surprise to us here at TalkTo. We make a mobile app (called “TalkTo) that allows anyone to text any business. If the business, is on our platform, great. If not, no worries; TalkTo will get you an answer, even if that means calling the business.

Lately, we’ve been getting a lot of our new customers via mobile Facebook ads. In fact, I’ve been called a liar when I tell people that we created a campaign that on a single day had an 11 percent click-through rate and a cost just $0.30 per install.

It took us awhile to figure out how to do that. I want to share with you six most important lessons we’ve learned along the way.

1. Men want to be a part of something big--anything big

We tested this ad on both women and men. Here’s our dirty secret: the bar chart isn’t based on any data at all. It just shows, using visuals instead of words, that something exciting is happening. When judging the effectiveness of our Facebook ads, we look at two metrics: click-through rate and, more important, cost per download, or the price we pay for each person who actually installs the app. 

For men, this has been one of our most successful ads, with a click-through rate of more than 3 percent and a cost of download of just 58 cents. But for women? We had a click-through rate of .145 percent, and a cost-per-install of $0.00. That’s because not a single woman downloaded the app. Women, apparently, are a bit more discerning than men.

2. Women want the what and why

Our mostly-male team was having trouble designing ads that appealed to women. This one did poorly. (Note: a woman in our office suggested this ad -- please don’t accuse me of using stereotypes!).

So I asked my wife: what do women want? Her response: “Women feel overwhelmed. At work, they’re just as busy as men. Then they come home, and they’re responsible for the upkeep of the house and the kids.” We Googled images of “busy women,” and a surprising number of them depicted women with lots of arms. So we decided to build an ad around that:

We think this ad succeeds because it’s both emotional and educational. It explains what TalkTo does (we make your phone calls so you don’t have to), while demonstrating why you should want it (it reduces stress). It has a click-through rate of more than 2.5 percent and a cost-per-download around $1.

Here’s another successful ad, which we aimed at women with children under the age of 4:

This had a click-through rate of more than 2 percent, and a cost-per-install of under $1. It also gets at the what (text millions of businesses) and the why (get stuff done without waking the baby).

Please note: it’s best to use original photos. First, stock photos look, well, stocky. Second, others (people, brands, organizations) have equal access to the image. It would be unfortunate if you shared an image with a less reputable advertiser. But when you need to move quickly and don’t have all the resources, stock photos are a great alternative.

3. Try small, cheap tests before launching a large campaign

We tried five different versions of the “Extra Arms” ad before we hit on the one that worked. The total cost of all the tests was only $45 because we set a limited budget for each one. 

Here are two of the failures:

Why didn’t these work? We can’t say for sure. It’s hard to predict which ones will be successful and which will completely flop. That’s why you should create small tests. Once you find something that resonates, roll it out to a larger audience.

4. Use Facebook ads to test your core messages

Here’s another version of the “Extra Arms” ad that didn’t perform well. It differs from our successful ad in only one way: instead of “offload your phone calls to us,” it says “text ANY business for free.”

This was a huge lesson: “text any business” has been our tagline ever since we launched. But for many people, “offload your phone calls to us” is a more powerful message. Based on this information, we’ve begun running more tests to see further refine the message. We haven’t reached a final conclusion, but Facebook provides a powerful medium for testing.

5. Clicks don’t necessarily lead to downloads

This ad had a click-through rate of almost 3 percent, but not one of those clickers actually downloaded the app. A clever slogan might attract attention, but you also have to explain your product.

6. Keep it short

This is our most successful ad so far, with a click-through rate of more than 4 percent and a cost-per-download of $0.43. It could be a testament to the power of dinosaurs. But in all likelihood, the ad works because it conveys a very clear value proposition (including what and why) in only six words.

Bonus: Mondays are your friend. Wednesday and Thursday, not so much

Costs vary depending on the day of the week. I reviewed eight campaigns that have been running for about five weeks. Based on average cost per install, Mondays are the best. Wednesdays and Thursdays were the most expensive. ad works because it conveys a very clear value proposition (including what and why) in only six words. There are thousands of factors that go into the cost per install--including demand from other adverstisers or a big, distracting event (anyone remember the royal baby?). The point is this: keep an eye on daily performance and adjust your spending for the best return.