At the digital media company where I work, we’ve dubbed this “the year of science.”
We’ve long relied on intuition and experience to create content that will engage readers and drive clicks to The Penny Hoarder. And while that has worked well, growing the site to 20 million readers a month, this year we’re putting more metrics and data behind our decisions. We believe that will go a long way toward helping us optimize, with the goal of continuing to grow our community.
For me, it’s also FUN. Using data to make decisions is one of my favorite parts of running the editorial team — it give me a whole new perspective.
Seeing things from a new perspective.
When I ran my own content-marketing shop, we worked super lean, investing only moderately in tools that would help us do the best work for our clients. At The Penny Hoarder, we’re doing much of the same work — but on a larger scale, with a bigger readership, a bigger company and a bigger budget.
While we still practice lean methodology in many ways, we have chosen to invest in tools that help us make data-based, real-time decisions about how to create meaningful content, reach more readers and bring in more revenue.
Here are three of my personal favorites.
(Use discount code MakeYourOwnLuck)
Naytev is a highly effective tool for growing your readership, both on Facebook and through website traffic.
I first met one of the founders, Patrick Costello, at a media event in Washington, D.C., where we both lived. He and the company have since moved to the West Coast, where they participated in an prestigious incubator program, Y Combinator. I tried the tool in their early days for client work, and then revisited it when I joined The Penny Hoarder — and I’m impressed with how the company has improved the product and narrowed their focus in the last few years.
Naytev helps publishers optimize content for Facebook through A/B testing. While publishers can use it in a number of different ways, at The Penny Hoarder, we use Naytev to test all our blog posts before sharing them on our Facebook page. That allows us to share an optimized version of each post, so we already have an idea of how it will perform once it hits the Facebook page.
As you might imagine, this is far more effective than posting on Facebook and hoping for the best. Using Naytev has had a huge positive effect on the size and engagement of our Facebook community, and it also helps us drive traffic back to our website.
Because they’re all tested before they go live, the majority of our Facebook posts get a ton of likes, comments and shares, which helps our content take off on other platforms as well.
How does Naytev actually work?
It allows you to A/B test three components of each Facebook post — the headline, image and teaser — to see which variation gets the most clicks and engagement.
Trying different combinations to find the one that performs the best is called A/B testing, but it’s actually more like A/B/C/D/E testing… you can try up to 16 variations. Those variations are shown to an audience that’s similar to your Facebook page fans, so you can see which one’s likely to do best on your page.
For example, here are all the variations we tried for a post we wrote about getting paid $18/hour to watch videos. (“Post text” is what Naytev calls the teaser, or the text you write above the photo when you share a Facebook update.)
(Click image to see a bigger version.)
That top variation was the winner, so Naytev showed us what it looks like in Facebook format:
Winning version in Naytev.
You can see it got an engagement rate of 3.87% during the test, and a CTR (that’s click-through rate, or what percentage of the people who saw the post clicked it) of 3.07%. Those numbers aren’t amazing, but they’re good enough to suggest this post will perform pretty well on our Facebook page.
(If you look closely at that screenshot of all the results, you’ll see some variations saw higher engagement and CTR than the winning variation. The one at the top is still our best choice because it performed best with a higher number of impressions.)
One thing I love about this platform is they make it easy to share the winning variation. You simply hit a button that says “post to Facebook,” and then you can schedule the post for any time, just like you would through Facebook’s interface.
Here’s what happened when we took that winning variation and posted in on our Facebook page. As we expected, it performed well. (That’s all organic traction; we didn’t pay to boost the post.)
Here’s the post on our Facebook page.
Here’s one big insight that will help you on Facebook even if you don’t use this tool.
While most people who A/B test look specifically at headlines, we find the component that makes the biggest difference in whether the Facebook post performs well is actually the teaser — that little bit of text you write above the image. It makes sense if you think about it, since that’s the first thing readers see.
While I’m a huge fan of this tool, Naytev does have shortcomings. Here are a few:
The monthly fee is based off how many tests you run, and its lowest tier is $359/month for up to 20 tests. (Patrick just let me know they also offer a plan for startups and bloggers at $189/month for 10 tests.) You’re allowed an unlimited number of users.
But you also have to pay for all the ads you run as part of the tests. Mind you, that money goes toward Facebook users seeing the ad and clicking over to your site, just like if you ran a Facebook ad through normal channels. But it does add up. Even if you spent just $25 per test, which is on the low end if you want to test a lot of variations, you’d spend $500/month for 20 tests in addition to the monthly fee. That brings the cost of the lowest tier to about $859/month.
Most digital media companies that use this tool to its full potential end up spending way more than that. But it’s so worth it for optimized results, and when you share the winning version on your Facebook page, it will likely hit a home run organically, without dollars behind it.
It takes a lot of time to create tests and track their performance. In fact, we just hired an employee who we expect will spend a third to half of her time testing our content. We believe it’s worth it.
A big reason for the time-suck is Naytev doesn’t offer an easy way to organize and track progress for a high volume of tests. I’m hoping this issue will resolve as the company matures, especially given how responsive the team has been so far to our requests for features.
It takes time to run tests, which means content can’t go live on Facebook right after you publish it. We’ve found ways to speed up testing, but since it has to go through Facebook’s ad channels, there’s no way to get an immediate result.
We write a lot of evergreen content, so we can take the time to test it. But publications that primarily share news that must be distributed as soon as it’s published would find less value in this system.
In some ways, this is a benefit: their customer support is awesome, and they’re responsive to our requests.
But the product still has some growing to do, and it will likely continue to change quickly as many young startups do.
If you want to try the tool, use code MakeYourOwnLuck to get 15% off your first month.
And even if you don’t end up using it, Naytev offers solid insights on their blog. They recently published a fascinating post with info I hadn’t seen elsewhere on the 45+ networks BuzzFeed uses to distribute its content.
This tool is a content marketer’s dream. When I first used it, a light bulb went off in my head: So THIS is how media companies keep close tabs on what everyone else is writing about!
CrowdTangle allows you to identify the best-performing content on your competitors’ social media channels, including Facebook.
Why is that amazing? Because knowing what performs well for others is a good indication of what could perform well for you. It’s an absolute wealth of ideas for blog posts and social-only content.
What CrowdTangle shows, to be completely honest, is kind of depressing for the journalism industry. The top-performing stories across the board are almost always about cats doing silly things, memes that make people laugh, and occasionally a celebrity death. This is what people share on social media, not meaningful journalism.
This, for example, is what came up at the top of my feed when I was writing this blog post. These are the most over-performing posts from the last 24 hours for all the outlets we follow, which includes plenty of hard news publications.
What CrowdTangle looks like.
CrowdTangle’s measure of over-performance is relative; it shows how well a post does compared to how posts usually perform on that Facebook page.
That means you can discover over-performing content even for Facebook pages that don’t have a huge audience. For content-creators, this is helpful because sometimes those stories haven’t been shared all over the web yet, and you could be the first big publication to write about it.
You can also search based on total interactions, which will give you results for over-performing posts from pages with big audiences.
For example, here’s what I see when I do a similar time frame as the first search, but for posts with the most total interactions:
What CrowdTangle looks like for total interactions.
Again, discouraging for journalists who are doing serious work. But even they can learn lessons about how to frame content from looking at what people tend to share.
While CrowdTangle has tons of magical features, my favorite is the ability to create lists of pages, so you can look specifically at competitors in your space to see what they’re doing well.
They also offer email alerts to help you stay on top of what’s trending, and Slack integrations so everyone on your team can know what’s happening without logging into email. And the BEST part is when you see your own page pop up in the platform with an over-performing post. Boom!
CrowdTangle costs $449/month for one platform, $899/month for three platforms or $1,299/month for five platforms (with an annual contract), though they don’t post prices on their website. With each platform you add to your account, you can have an unlimited number of users.
It’s available for Facebook, Twitter, Vine, Instagram, Reddit and YouTube, and the company says it’s working on a version for Pinterest.
We started with the Facebook platform alone, because that’s where we have the most followers. But we quickly added two additional channels: Reddit and Twitter, both of which are gold mines for blog fodder.
I don’t have any specific limitations to point out for this tool. That’s partly because I don’t use it on a daily basis — my team does — and partly because every time I’ve emailed them about a feature, our rep has shown me how to do exactly what I want.
CrowdTangle, like all the tools on this list, offers fabulous and personalized customer support. Not only did they give us a great demo that showed how we could use the platform for our specific needs before committing, they paired us with a rep who who’s easy to reach and helps us with anything we need.
Parse.ly is an analytics platform specifically for digital publishers. It shows how your content performs, so you can use that information to create content that will perform even better in the future.
Sounds a lot like a free tool many of us already use, right? Google Analytics.
I thought the same thing before I demo’ed the platform. While we already used Google Analytics heavily, we found Parse.ly offers a number of features you don’t get with GA.
Here’s what makes it different:
Google Analytics can be challenging to use, especially if you’re not familiar with the platform. It takes a while to figure out where to find everything, and it’s not easy to set up reports or add filters to nail down the information you want.
Because of that, editorial staffers might shy away from logging in at all. In fact, a lot of companies have a tech person everyone goes to when they need help finding data in GA.
We wanted to foster a company culture where everyone’s decisions are data-based — and that means making it easy for every staffer, no matter their role, to find the data they need.
This is where Parse.ly shines. The software is easy to navigate, even for first-time users. Even so, our rep was awesome about giving several groups of staffers personalized tutorials so they’d feel comfortable diving in.
Here’s what Parse.ly looks like when I log in:
What Parse.ly looks like (click for bigger image)
It gives you the big picture, while also allowing you to dig deep into specific posts, authors, sections, tags and referrers to see what’s working.
This sounds trivial, but the way Parse.ly turns data into visualizations helps the user better understand what it means for their work.
In Google Analytics, you get a bunch of stats but not many graphs. You can see from the screenshot above that Parse.ly is much easier on the eyes. Not only is it more appealing to look at, it’s easier to find what you’re looking for and understand it.
Since Parse.ly caters to digital media companies, it offers features you simply won’t find in Google Analytics. One feature — and this is the feature that sold me on using their platform — is author reports.
Parse.ly allows you to easily group posts by author, so you can see how an author’s posts have performed over the last week, month or year. You can also compare metrics from one author to another, so you can see, for example, which of your writers brought in the most page views in the last month. You might even use that information to evaluate freelance writers, so you can solicit more content from high performers.
Here’s what Parse.ly shows for a specific author. These are recent stats for one of our staff writers, Susan Shain. You can see how the posts she’s written have performed over a certain time period, how that compares to a previous time period, and which of her posts performed best during the selected time period.
(Click image to see a bigger version.)
Parse.ly even allows you to share links to dynamic reports with freelancers who aren’t on staff and don’t have a login for the system, so they can track their post performance, too. I could see this being incredibly useful for publishers that offer a bonus system for freelancers based on page views.
One important feature that’s lacking from Google Analytics is easy-to-create reports.
That’s a core feature on Parse.ly; you can automatically generate one-time or recurring reports that show any stats you’re interested in. Email those reports to yourself, to your colleagues, or, if you have API access (which costs an additional fee), post them directly into Slack.
If you have screens around your newsroom like we do, Parse.ly goes into full-screen mode with the click of a button.
That lets you display whatever stats you think are important, including real-time information about how your content is performing at this very moment, to your entire team.
We find rallying around metrics and data-based goals makes work more fun and boosts camaraderie.
The Penny Hoarder Office in St. Petersburg, Florida
The tool does have its limitations, so there are still some stats we pull from Google Analytics instead.
The big one is the inability to separate paid traffic from organic traffic. We do that via Google Analytics, allowing our editorial team to focus on organic traffic, so we know how much traction posts get without putting money behind them. Parse.ly says they’re working on this feature.
Parse.ly doesn’t make their pricing plans public, but we pay $3,000/month for an unlimited number of users. If you sign up here and mention my name when you do, they’ll give you a 10% discount.
As I mention in the headline of this post, big media companies are using these tools.
All three of them are pretty pricey for your average blogger. If that’s you, here’s what I’d suggest, both for these tools and for others you’d like to use to grow your readership.
Some companies, especially those that are relatively new and still building their user base, might be willing to offer a deal for smaller shops as a way to get user feedback and buzz about their product.
Even if the fees listed are out of your reach, it doesn’t hurt to email the company and ask if they offer any discounts for bloggers or startups. You’d be surprised how many shops will be happy to help, especially if you might help spread the word about their product to your community or network.
Sometimes a tool that seems prohibitively expensive becomes doable when you do the math on how much it would help you earn. Spending $100 a month on a tool might sound like a lot, but if it helps you earn an additional $200 a month, you’re coming out ahead on profit, while continuing to build your brand.
Of course, if you start investing in a lot of pricey tools, those expenses quickly add up — and when you’re a small business owner, you should aim to keep expenses low. But while not spending money is smart in most situations, sometimes it literally pays to make an exception if you see a solid return on your investment. (Here’s another example of when it might make sense to spend as a small business.)
Even if it doesn’t make sense to spend on tools like this, you can learn better ways to grow your business simply by knowing what successful digital media companies are doing. There’s often a cheaper, more time- or effort-intensive way to get similar results without the tool.
For example, if you don’t want to pay Naytev’s monthly fee, you can still A/B test ads through Facebook’s ads manager, it’s just a bit more clunky and takes more time.
If you love data but can’t afford Parsely, Google Analytics offers many of the same features — you just might have to put some time into learning how to use it. And you could learn from your competitors on Facebook without CrowdTangle simply by visiting their pages to see what gets a lot of interaction.
Got questions on any of these tools? I’m happy to answer them in the comments.
Disclaimer: I added affiliate links above for companies that have referral programs. That means I earn when you sign up through my link.
These are great tools! I have one more for A/B testing that is super good too…Optimizely. I highly recommend it. There’s a couple more small tools I’ve used before like Grammarly (for grammar), Skitch (for making tutorial pictures), and Scrivener (a great tool for writers but I haven’t used it myself). Great post!
Thanks for sharing these tools and breaking down how they can add value. Love the concept of data-driven content but having trouble implementing it! This will help 😀
These look like some great tools to grow your digital company. While they can be pricey, it is important to consider the ROI as you mentioned, and often times it can pay off big time! Thanks for sharing these resources.
We are a publisher that recently started testing paid media on Facebook with Naytev and are struggling to get our CPCs down. We average $0.50, but have seen as low as $0.20. Where should we focus our efforts in trying to lower the CPC? Ideally we’d like to be closer to the numbers TPH is showing in this article <$0.10. Does it mostly come down to target audience and ad relevance?
Hi Jasmine,
Yes, it does come down to audience and relevance. A target audience that works well for us — if you have a big enough Facebook presence to support it — is to create a look-alike audience that’s similar to your Facebook fans. Naytev will likely also help you hone in on an effective audience if you ask them for help.
Great post. I was thinking about trying some minor A/B testing with genie.us but it looks like Naytev might be a more versatile option. Thanks!