by Tim Staples
Once you have that information, you can use the following questions to learn as much as you can and spot patterns. Overall, how is the quality of content in the space? Which brands have been most effective? Which brands have been least effective? Of the effective brands, is there a consistent theme or voice to their content? What topics are resonating with their audience? Which topics are getting the lowest response? Is there a format that is consistently performing the best? Are videos working better than photos on certain platforms? How often are they posting? Is there anything you can learn from the time of week and even the time of day the successful posts went live? What is the engagement rate of your competitors? What in the category as a whole is doing well? What in the category is not doing well? Where is the content opportunity that nobody is owning?
If you take the time to do this, it will give you invaluable feedback as to where the opportunity is and what audiences in your specific space are responding to. It will also help position you in relationship to your competition so you can carve out a niche.
Bucket 3: Dream Big
The third bucket is the dream bucket. This entails all the brands or personalities that you aspire to become over time, your digital idols that are blazing the trail and having massive success with their social content. It’s a great exercise to define who you admire in the space and who you dream to be, so that you can start to build the roadmap to get there.
As with the competitor bucket, I’d suggest choosing five to ten brands or personalities that you view as being highly successful with their content. Your list can include both brands and personalities, and they don’t need to fall within the exact space that you are playing in. But they should include a few examples that have some level of industry relevance. Where this differs is in the approach to how you study them.
First, spend some time poring over all of their content. Take a look at their early posts, from back when they were just starting out on social media, and then compare those videos to what they are putting out now. What has changed? What have they learned, and how have they adjusted?
In most cases the early content will appear much less focused and “in the zone” than the current content. This is good news! Even your heroes took time to find out what they do best.
Now look at the top ten videos or posts for each of the examples, and also analyze the ten that had the least engagement. What jumps out at you between the two categories? What did they do right when their content succeeded? What do you think they did wrong when it didn’t work?
After you’ve got a good feel for all of their content, ask the following questions: What makes each brand or personality special? What is their unique voice? How would you explain them to a friend, in one sentence? What topics do they cover? What trends are they riding? What value do they provide to their audience? Do they flip the script? How approachable or distant do they feel? What are they doing better than anyone else? What could they be doing even better?
Answer these questions and you’ll be a big step closer to where you want to be.
Level Two: The Content Test
Once you have done all the soft testing you can, it’s time to take a brave step into the real world—you need to test your content on a real audience.
This sounds scary, and many brands are very nervous about the idea of making something go live before they are completely satisfied with it, but it’s a key part of success on social media.
Because the algorithms that control our social platforms are so vast, and their computing power so great, tapping into them early and running your content against them to see how it performs is absolutely crucial to ensuring success. This is especially true because those algorithms are the AI gatekeepers to your audience. In short, if they don’t like your content, either your audience won’t see it or, best case, you will have to pay a premium for the privilege. For example, the cost of a view on Facebook for a traditional ad that is lower in the algorithm may be 6 to 8 cents. Meanwhile, the cost of a paid view for a shareable piece of content may be only 1 to 2 cents. That is a huge difference!
To explain the importance of this, let me share another one of our spectacular failures. When creating a campaign for FitTea, the herbal supplement beverage to help an upset stomach, we decided to ride the wave of a big movie coming out, called The Purge. We were right to bank on this trend, as the movie franchise has proven hugely successful and spawned numerous sequels. We created a killer piece of content, a complete spoof of the trailer for The Purge, and designed to launch it days before Thanksgiving. The entire horror premise was built around the idea that people overeat during the holiday. Whereas in the movie, all crime is legal for 24 hours, in our trailer, all food was edible for 24 hours. People literally roamed the streets like zombies, looking for turkey feasts, breaking into stores, and raiding candy aisles with ravenous fervor. And the title was the real kicker. A direct play on the title The Purge, our fake movie was called The Binge.
Everything was perfect. Except the timeline. Not only did we not have time to do any tests, but the job was awarded so late that we finished the video hours before it was meant to launch. We didn’t even have time to run it past the ad platforms to see if it passed . . .
It didn’t.
The ad blockers thought that there was disturbing content in the video. There really wasn’t—our “zombies” were just people eating turkey legs—but that didn’t stop the automated visual programming from flagging it and forbidding any spend. Because it got flagged, the video got pushed down in the algorithm, and we had no way to put a single promotional dollar behind it. All of that glorious creative work we were all so proud of just launched weakly and languished. It launched from a relatively dead page, with very limited engagement or traction.
The original idea was that our campaign would change all that, that we would put some spend behind the content to kick-start it, and then the internet would do its thing. But without a push, it never gained any traction. The campaign lingered with hundreds of thousands of views instead of millions—all because we didn’t have time to test it.
Now, we know better than to take on these tight turnarounds.
Whenever we can, we bake a two- to four-week testing period into our work on any campaign. We then spend that time creating endless variations of the video and running them as “dark” posts, to see what resonates with certain audiences.
There are a lot of concepts to understand here. I’m not going to cover how to buy paid media on Facebook; if you’re not familiar with it, there are countless articles on the web that can walk you through it. So let’s start with the idea of “dark” posts. A dark post is something that you publish but that does not show up in your timeline. It’s a setting within the Facebook ad platform, for example, where your post gets pushed to a real audience exactly the way a “real” or “live” post would, meaning it shows up in their feed and they interact with it exactly the way they normally would, but the difference is that it doesn’t show up in your feed. That means that no one can search for it or find it by looking you up. The post is “dark” to anyone except for the people you are pushing it to.
So what do you test with dark posts? Well, trying out several different versions of the beginning of your video is a great place to start. You know the first 7 seconds is everything, so test different versions of that. Which opening image gets people’s attention the most? What opening text copy makes more people go full screen? Does the color of the text matter? Does the animation in or around the text have an impact? What if you start the story at the end and create a teaser before you go into the real thing?
There are a million variables, and the key is finding those that truly matter to the story you are telling. This changes on every video, but it often connects to the very beginning.
That said, seeing how the data plays out over time is also very interesting. When we did our photo restoration project for Adobe, for example, we realized that even though a particular edit of the video tested slig
htly better for initial viewership, there was another version of the video that came in a close second on viewership but actually had slightly higher retention through the last quarter of the video. In particular it carried more people across the messaging at the very end, telling us that the desired impact of the video would actually be greater. In that case, we went with the second option.
It’s also important to understand what you are actually testing for. While it may sound like you are testing for creative feedback, to see what people think of your content, and while that can be true, more often than not, you are really testing for what the algorithm likes. You are getting data back that says the algorithm will serve this particular piece to more people, or for a lower cost. That is all well and good, but you have to keep the effectiveness of your campaign in mind and remember the ultimate mission. It’s easy to get lost in reaching for the most views every time, but at the end of the day, is that what truly matters? Probably not.
Level Three: The Audience Test
Now that you have your content honed, and you have a sense of what the algorithms like, it’s time to focus on your audience.
The tools for segmenting your audience on these platforms are insanely sophisticated. You can divide people not just by basics, such as age, gender, ethnicity, and geographic location, but also by whether or not they like Alice Cooper, dogs, or bicycle pumps.
As part of our learning process we launched a channel called Like It GRL, which was geared squarely to teenage girls. It was an exercise in seeing how effective we could get with our targeting. We spent months tweaking details, not even video content but post copy, just to figure out what kind of language we needed to speak, to reach our desired demo. We learned things that may seem obvious and intuitive, such as teenage girls absolutely love emojis and generally don’t want to read too much. Having the actual data to back this up gave us the confidence to make it a rule that our posts were to include massive amounts of emojis but no more than three to five words. Our audience skyrocketed and our costs went way down.
The way we do this with our bigger campaigns is very varied, but the philosophy is always consistent. We start by figuring out our ideal target audience. Sometimes this comes with great specificity from our clients, telling us they want to reach a very precise group of people, based on some much larger objective they have. They might be running a campaign for back to school, and they want to reach only married parents of two or more children in middle-class households in the coastal regions, for example. Or an even more specific campaign, like our first video for mall company Macerich, asking us to focus our campaign for the King’s Plaza Shopping Center in Brooklyn around the geographic region where the Town Center (Macerich-speak for “mall” ) was located. We drove a million views from within a few miles of that mall alone. If you know someone in Brooklyn, ask them if they have seen it. Odds are they have.
As you test your content, you will see that different iterations speak to different audience groups. This will teach you how to hone and refine not just your content, but also your audience. Perhaps there is some metric that is not hyper-relevant to your messaging but that will have a huge impact on your spend because the algorithm is much happier serving the content to that particular subgroup. Without testing, there is really no way of knowing.
Now, the cost.
Well, running the tests doesn’t have to be very expensive at all. The amount of money you spend really only dictates two things: how many tests you can run—and how wide an audience they will reach.
The way this works is that you create anywhere from two to five different versions that you are looking to run against each other. Five is the maximum because past that, it gets really complicated to analyze and the variables are too great. You load them into your ad account, set the spend you are willing to put against them, and hit run. (Okay, that’s an oversimplification, but essentially right.) The spend can be anywhere from a few dollars to a few hundred to a few thousand and up, but that’s all you need. Even the smaller budgets will give you an audience size of thousands of people, certainly more than any TV focus group ever.
The trouble is time. Each iteration needs to run for at least two days. If you are testing 10 or 20 different versions, well, do the math . . .
At Shareability, we do all of this using software, but unless you have access to some costly subscription-based or proprietary software, this is a labor-intensive task. We might be able to do it in a few weeks, sometimes even a few days, but do not underestimate the effort involved. Learn from our mistakes and make sure you have enough time on your hands to run all the tests you can imagine.
A Final Note: With Paid Media, All Content Is Not Created Equal
When it comes to paid advertising on social media, all content is not created equal. On television, when a company runs a TV commercial on a Thursday night, that spot is priced based on the time slot. Taco Bell and Ford pay the same $800,000 to run a 30-second ad, and that ad receives roughly the same exposure, no matter what commercial they run. The internet is completely different, as advertising on all the major social platforms is auction-based, meaning the price is determined by how easy or difficult it is for the platform to serve the ad. And because social media is extremely targeted, you don’t need millions of dollars to advertise. In fact you can start as low as a few dollars to see what results you initially get and then spend increasing amounts when you see that an ad is working. A few thousand dollars can have a big impact with a targeted demographic.
One of the most important features in determining this on Facebook, one that a surprising amount of people in the advertising world don’t even know about, is the Facebook relevancy score. For every piece of promotable content uploaded to Facebook, the algorithm assigns a relevancy score of 1 (worst) to 10 (best) based on a number of factors. Most important is how viewable the algorithm considers the content to be, based on the engagement of the initial audience that views the videos. In truth, most advertisements provide little to no value to the audience and therefore have extremely low engagement, which in turn results in a low relevancy score, in the 1–4 range. Quality content that is marginally shareable falls into the middling 4–6 range. And highly shareable content, which receives substantial engagement, receives a relevancy score of 7–10. This score is everything when you are putting paid promotion behind your content. Videos with high relevancy scores climb the algorithm, more people want to see the content, and Facebook in turn charges you less to push paid views. On the opposite end, content with low relevancy scores drops down in the algorithm, making it harder for Facebook to distribute that content, and therefore the price goes through the roof.
By posting a brand video that drives substantial engagement, you receive a higher relevancy score and trigger the algorithm to push that video out to a larger audience at a much lower cost per view. Conversely, if your video receives minimal engagement, it will slide down in the algorithm and you will have to pay higher rates to push it out.
Let’s say a cereal company runs a dull ad about the fiber content of its bran flakes. An ad with such minimal shareable value to the viewer will result in predictably low engagement rates, and therefore it will receive a relevancy score of 1 or 2. The lower relevancy score causes the Facebook algorithm to work harder to deliver the ad, and therefore Facebook needs to charge more to push it out. A score of 2 might mean that the cost of a view is in the range of 8–10 cents. Contrast that with one of our hero videos that is rapidly viewed, liked, and shared by hundreds of thousands, and consequently stamped with a relevancy score of 9. Because that video is being proactively viewed and shared, the Facebook algorithm actually becomes our ally and helps push it out to a broader audience and charges far less money to do it, sometimes at less than 1 cent per view.
This dynamic presents a huge opportunity for brands that are funneling their media budgets away from traditional media like television and into social and digital media where—if they follow the principles outlined in this book—
they can break through the noise, drive massive attention, build deep relationships with customers online, and save millions of dollars in the process.
Epilogue
Finding Your Way with the New Rules
I hope you enjoyed reading this book as much as I’ve enjoyed writing it. That’s because this exercise provided me with an opportunity to take a breath and look back over my career, and to reflect on how much has changed. Over the past two decades, technology has changed the way in which human beings communicate, it has changed the way we interact with each other, and it has revolutionized the way in which we interact with brands and with new personalities on the rise. The digital revolution has disrupted virtually every brand category, from Uber to Netflix to Dollar Shave Club, and in addition, personalities and celebrities have come forth, all due to the unprecedented impact of social media. Any time there is disruption at this level, there is also a tremendous opportunity. An opportunity to embrace the amazing tools that digital gives you. An opportunity to use content and data to build relationships directly with your customers. An opportunity to build a brand that is more valuable than you ever thought was possible.
My favorite part of my job is noticing how fast everything changes. My partner Nick Reed and I often joke that we really don’t want to leave the office at night because when we show up the next day, things will be different. We’ve found that when you actively embrace that kind of change, it can take you to some pretty fun places.
