First-of-its-kind data analysis of sponsored content shows that it’s better to mention brands strategically rather than repeatedly.
It’s a question that brands and publishers have pondered for a long time: how and when should a brand be integrated into sponsored content?
Thanks to a new study of over 300 pieces of sponsored content, we finally have the data to answer it.
Now, we’re not talking about whether or not to disclose a brand association. At Pressboard, transparency is a core company value and therefore a mandatory requirement of every single article that is created and published through our marketplace. No, what we’re talking about is that tricky process of mentioning brands and products in an article itself in a way that doesn’t turn off readers.
We all know how quickly readers can figure out whether a brand is integrated poorly. We see it in the comments section, or in the amount of time readers spend on the page. So, we turned to the experts (the readers) to see if we could learn a thing or two from them.
Pressboard collects tens of thousands of data points every day from the sponsored content created through our platform, most of which is related to how readers are interacting with the content. To understand the impact of brand mentions in sponsored content, we dug through that data to answer the following questions:
The results finally prove what content experts’ instincts have been telling them all along:
My hope for this research is that it encourages advertisers to boldly but thoughtfully integrate their brand into the sponsored content they create with publishers, because the data highlights that it can be done well! – Jerrid Grimm, Co-founder, Pressboard
The study also highlighted a few insights that we didn’t expect.
As it turns out, sponsored content in which brand mentions are placed sparingly and strategically actually performs better than sponsored content devoid of any brand mentions. Many in the industry have often assumed otherwise: that sponsored content performs better when the brand isn’t mentioned at all. However, it appears that a thoughtfully placed brand mention can actually engage readers and encourage them to read more of the content.
Sponsored articles without any brand mentions performed well (with an average active reading time of 63.5 seconds and a scroll rate of 78%), but still failed to outpace content with one brand mention.
Two words here are key: sparingly and strategically. Where and how many times the brand is mentioned can have a significant impact on the effectiveness of sponsored content, so it’s important to get it right.
To help publishers and their creative teams produce more impactful content, we’ve outlined the effect that placement and occurrence of brand mentions have on reader engagement.
For this analysis, we considered two factors: placement (where the first brand mention occurs in the sponsored article) and occurrence (the number of brand mentions in a sponsored article). We then analysed their impact on two metrics:
Our sample comprised 331 sponsored content pieces that were published and measured through the Pressboard platform by North America’s top brands and publishers. To ensure the statistics were consistent, we normalized the content to calculate reading time and scrolling behavior as if the length of each article was 750 words. You can read more about this in our methodology section.
When the first brand mention occurred later in a sponsored story, readers were likely to spend more time engaging with the content.
Articles in which the brand was mentioned in the first 100 words saw an average active reading time of 56.2 seconds. Compare this to articles in which the brand was mentioned after the first 300 words, which saw an average active reading time of 68.1 seconds. That’s a difference of 12 seconds that a reader is spending engaging with (or not engaging with) the content.
This essentially gives us a sweet spot for brand mentions, letting publishers thoughtfully integrate a brand while also improving the overall performance of the story.
However, there are limits to how late in the article a brand should be mentioned. The results show that when the first brand mention appeared near the bottom of a sponsored piece, in the 600-750-word range, active reading time was reduced, falling to 66.3 seconds.
Sweet Spot: 300-600-word range
Where a brand mention was placed had a slightly less dramatic effect on the percentage of the content a user scrolled through. Still, the results are in line with active reading time.
Data shows that if the brand mention happened in the first 100 words of the piece, readers only scrolled through an average of 72% of the article.
When the first brand mention occurred in the 300-600-word range, readers scrolled through 80.5% of the article. Scrolling behavior continued to improve when the brand mention occurred late in the article, with readers making their way down an average of 81.5% of the page.
This is interesting because while a late mention negatively affected active reading time, it had the opposite effect on scrolling behavior. A reader will still make it most of the way down the page, regardless of where the brand is placed.
Sweet Spot: 300-600-word range
The most dramatic result we saw concerns the effect the number of brand mentions had on active reading time.
When a brand was mentioned only once, readers spent an average of 69.6 seconds reading the article. As soon as another mention occurred, that number dropped to 61 seconds, and as more brand mentions were added, reading time continued to fall. Readers spent an average of 57.3 seconds with articles with three to four brand mentions, and 55.5 seconds with articles with five or more.
This is consistent with most publishers’ expectations. If a reader feels that the focus of the article is more on the brand than on the story that’s being told, they’re less likely to engage with it.
Sweet Spot: 1 brand mention
In branded articles where the brand mention occurred only once, readers made it an average of 80% of the way through the article.
As brand mentions increased, the percentage of the page that readers scrolled through fell: in articles with two brand mentions, readers scrolled down 76.8% on average; in articles with three to four brand mentions, the number was 76.4%; articles with more than five brand mentions saw the lowest scroll rate, of 74%.
Though there is an evident relationship between occurrence and percent scrolled from examining the data, the relationship is not statistically significant. With more data, it is likely that we’d see a strong relationship develop.
Sweet Spot: 1 brand mention
In reviewing the data, something that surprised us was that leaving the brand out of the article entirely doesn’t result in more engaging content.
Branded content that mentioned the brand in the 300-750-word range had a higher active reading time than content that didn’t mention the brand. Readers also scrolled further down the page when the brand appeared after the first 100 words than they did when no brand was mentioned.
Sponsored articles without any brand mentions performed well (with an average active reading time of 63.6 seconds and a scroll rate of 77.9%), but still failed to outpace content with one brand mention.
This is a good lesson to publishers and creatives that you don’t have to hide or bury the brand when creating sponsored content. By being strategic about where and how it’s being mentioned, you can actually drive curiosity with readers and encourage them to engage more with your content.
We always urge brands, publishers and agencies to be completely clear about who is behind a piece of sponsored content. This great research from Pressboard proves our point beautifully. – Jesper Laursen, CEO, Native Advertising Institute
The findings related to both active reading time and percent scrolled support the conclusion that brands should be mentioned sparingly and strategically in branded content. With that in mind, the study produced important insights that should help guide your content creation efforts.
Q: How early in a piece of sponsored content should a brand be mentioned?
A: The best place to first mention the brand is part-way through the article, in the 300-600-word range. By introducing the brand after the introduction, it enables the brand to add value or insights in a supportive role, as opposed to being positioned as the focus of the piece, which may not be as interesting for readers.
That said, brand integration shouldn’t be saved until the end of the article. Doing this may cause the audience to exit due to perception of having been tricked into reading to the end of a story that turns out to be branded.
If the product messaging doesn’t naturally fit within the article itself, it might be better to include it in a branded footer at the end of the content. Every sponsored article published through our platform included a branded footer that we encourage publishers to use to include key brand messaging and a strong call to action.
Q: How many times should a brand be mentioned in a piece of sponsored content?
A: When it comes to deciding how many times a brand should be mentioned in an article, the rule is simple: less is more. Mentioning the brand doesn’t penalize you – in fact, the data suggests that it’s better to introduce the brand than to avoid including it at all. But publishers and their content partners should do so sparingly, as with each additional mention, readers lose interest and disengage.
“Our research paints a pretty clear picture of the best practices of sponsored content. Consumers really want brands to clearly disclose their sponsorship of sponsored content, featuring both their name, logo, and a clear labeling like “advertisement” or “sponsored”, but they don’t want to be sold to. Self-promotion in branded content kills trust—by 27%. Tell stories people crave. Don’t try to dress up a display ad in sheep’s clothing.”
Joe Lazauskas, Head of Content Strategy, Contently
“Since mentioning the brand results in greater performance than ignoring it completely, and mentioning it mid-way creates better engagement than waiting until the very last second, I hope that this data will help lay to rest the idea that a brand would benefit from ‘tricking’ the audience into believing branded content is editorial for all or a majority of the piece. Brands absolutely have a right and responsibility to be present in their content, but it has to be done carefully. And if a brand feels like their content requires a brand mention in the first hundred words, its likely that the content concept needs to be revisited with a critical eye for whether that content is truly in service of the readers or only in service of the brand themselves.”
Melanie Deziel, Branded Content Consultant & Speaker, mdeziel.com
“For years we’ve been educating advertisers about strategically and sparingly mentioning their brand in sponsored content. Now, the industry finally has the data to validate our intuition. My hope for this research is that it encourages advertisers to boldly but thoughtfully integrate their brand into the sponsored content they create with publishers, because the data highlights that it can be done well!”
Jerrid Grimm, Co-founder, Pressboard
“Transparency is one of the most debated topics in the native advertising industry, and for good reason. We always urge brands, publishers and agencies to be completely clear about who is behind a piece of sponsored content. For legal reasons, for moral reasons, but also because we truly believe that if you publish relevant, valuable and engaging content, the audience is perfectly fine with listening to a brand – as long as you do not overdo it. This great research from Pressboard proves our point beautifully.”
Jesper Laursen, CEO, Native Advertising Institute
“I’ve always said that great content can come from anyone, including brands that ultimately want you to buy their product or service. But an in-your-face approach can backfire. What it really comes down to is trust and it takes some time in a piece of content to establish that. Is the brand helping solve a business problem? Is it helping validate or spur creative new ideas? Is it entertaining the reader? If the approach is genuine, there is greater value in a brand mention. On the other end of the spectrum, brands that try too hard to hide can come across as deceptive – and the penalties are huge.”
Stephanie Stahl, General Manager, Content Marketing Institute
“I’ve often talked to brands about the “slider” approach to brand mentions in sponsored content. There’s room to move the volume slider left and right, but slide it to far in one direction and you’ll turn off readers in a hurry. I am particularly intrigued by the stats on WHERE a client first appears in content.”
Jordan Hyman, VP Business Development, The Wall Street Journal Custom Studios
One challenge this study presented was ensuring that we captured every possible instance where the sponsoring brand was mentioned in a piece of sponsored content. With tens of thousands of data points to analyze, manually verifying each client mention would have taken us years to accomplish!
To speed up the process, we used a natural language processing system to identify named entities within the text and count any entities that were similar to the brand’s name. The similarities between these named entities and the brand names was calculated using Jaro-Winkler similarity measurement. All entities that had a similarity above a prespecified cutoff (0.7) were considered brand mentions. For example, phrases like “TD Investments” or “TD Line of Credit” were counted as a mention of the brand TD.
We excluded articles from the analysis where the sponsored content was not about the brand directly. For example, articles sponsored by tourism boards, which often promote businesses and events in the region, were not included in the study.
The amount of time readers spend engaging with branded content is obviously impacted by the length of the content. To make sure that stories of different lengths were analyzed equally in the study, we used a normalization method that divided the average time spent on page by the number of words in the article and then multiplied by 750. That calculation computes the average time of engagement for every piece of sponsored content as if the length was 750 words. We also limited the analysis to stories between 250-1000 words in length.
Statistical analysis was run using SciPy and NumPy. The relationships were tested using a One-Way ANOVA test. The F-statistics calculated were under the null hypothesis that the group means are equal. The alternative hypothesis was that at least one group is not equal.
Natural Language Processing package: spaCy available from www.spacy.io.
Story content and meta-data were scraped using this python package.