HockeyStack’s Self-Reported Attribution Report – 2024
We pulled over 8,000 self-reported attribution responses from high-intent prospects across dozens of B2B brands globally, to unravel patterns, preferences, and peculiarities of their Self-Reported Attribution answers.
All HockeyStack Labs reports are done using anonymized HockeyStack customer data. We did not partner with anyone on the creation of this report and it was not sponsored by a vendor.
Understanding how customers discover your brand is hard.
With so much complexity in the buyer journey, B2B brands have had to find creative ways to measure marketing and triangulate the truth about what works and what doesn’t.
Enter: self-reported attribution (SRA), or asking prospects how they heard about you.
Over the last few years, SRA has re-emerged as a radically simplified way to discover how buyers discover you.
But how does self-reported attribution stack up against other measurement methods?
How reliable is SRA as a dataset?
And what can we learn about today’s B2B buyer journey from SRA alone?
Since HockeyStack tracks self-reported attribution for all of our customers, we decided to find out.
How?
We pulled over 8,000 self-reported attribution responses from high-intent prospects across dozens of B2B brands globally.
The goal?
To unravel the patterns, preferences, and peculiarities of SRA, including:
- Why podcasts and Slack communities get mentioned so infrequently?
- How demand capture still dominates marketing motions?
- What channels drive the worst traffic (ahem, display)
- Where last-touch and SRA overlap the most?
- And a ton more…
Let’s jump in!
What is self-reported attribution?
First, what is self-reported attribution for those who may not know?
Self-reported attribution (SRA) is a form of measurement through surveying. And it’s one of the oldest forms of customer surveying in marketing.
It’s pretty simple: on your demo form (or in a sales call), you ask prospects how they heard about you- hence the term “self-reported.”
That’s it.
Self-reported attribution is neither correlative or causal, meaning it doesn’t seek to estimate incrementality (lift) or marketing influence. It’s merely an attempt to curate directional signals about what’s working and what’s not, and to do it fast.
Over the last few years, SRA has grown in popularity within B2B marketing as a way to gut check hard-to-track brand and demand activity like social media, podcasts, and content.
What’s the purpose of this report?
This report is part one of a deeper analysis on self-reported attribution.
In this report, we wanted to kick things off with a high-level overview of self-reported attribution responses, their most common categories, patterns we discovered in the data, and compare it with the last touch source from HockeyStack.
Previous research on self-reported attribution narrows in on a small dataset, usually the company’s own data. We wanted to do something different and pull a more representative sample from broader B2B.
Since self-reported attribution is in large part a reflection of your go-to-market motion, biasing the data with a limited dataset doesn’t tell us much about the state of go-to-market in B2B.
This report will give you a glimpse into the efficacy of self-reported attribution, along with a peak into what channels B2B buyers discover brands through most often.
Where did we get the data?
We pulled over 8,000 self-reported attribution responses from dozens of consenting companies with firmographics ranging from:
- Industries: All in B2B mostly Software
- Regions: North America, Europe, and Australia
- Company size: Mid-market and Enterprise
All SRA responses came from high-intent hand raisers who booked a demo through the website’s main “contact sales” or “book a demo” form.
Between all the forms, only two used a drop-down menu with pre-defined SRA choices while the rest used an open field text box (recommended).
Last, forms varied from multi-step and single row to double row and Calendly pop-ups.
Let’s explore the findings.
The Self-Reported Attribution Report
Of the 8,528 self-reported responses, search engines took home the lion’s share, followed by social media, word of mouth, online, and content.
Surprised? Probably not.
Since self-reported attribution is a reflection of your go-to-market motion, it makes sense that search would lead, followed by social media and word of mouth.
But let’s dive deeper.
Search reigns supreme
Raking in nearly half of all self-reported mentions (45%), search engines take the number one spot by a landslide.
Search includes all explicit mentions of search engines like “search,” “keyword search,” “Google,” “Google ad,” and “Bing.”
Unsurprisingly, 38% (1,449) of all search SRA responses mentioned Google specifically. But it was likely more than that: at least some of the 683 generic responses like “internet,” “web,” and “online research” (grouped under “Online” in this report) likely came from search engines as well.
What does this say about the state of B2B marketing?
First, search engines still dominate as a point of discovery for many buyers, and marketers can’t ignore it- no matter what rumors you’ve heard around B2B.
Second, it underscores the fact that a giant portion of potential buyers likely don’t know these brands exist when they move in-market.
If 45% of all high-intent demo bookers first heard about you through a search engine, they didn’t discover you passively while consuming brand media or social media. Either that or the demand generation programs some of these companies have run are more focused on narrow segments vs. market penetration- hence the lack of awareness in advance of purchase.
Could those responses have come from branded search, in which case they would have known about you before they began their search?
Unlikely.
“How did you hear about us” doesn’t ask “how did you find my website right now.” It asks for the point of discovery, which by definition wouldn’t have been a branded search.
And last, since SRA responses are a reflection of your go-to-market motion, this data suggests that B2B brands still rely heavily on demand capture over demand generation.
Linkedin gains momentum
Social media ranked second overall with 1,670 mentions, making up 20% of total respondents.
Social media included responses like “TikTok,” “Linkedin,” “Linkedin ad,” and “social media.”
Of those 1,670 responses, 30% (508) mentioned Linkedin, 3% mentioned YouTube (58), 2% mentioned TikTok (31), 1.3% mentioned Facebook (22), and less than 1% mentioned Twitter (9).
With 508 responses, Linkedin by itself has enough to rank #5 on the overall list, which should come as no surprise considering the B2B propensity of the platform.
Word of mouth takes third
Word of mouth took third place with 1,544 total responses (18%).
Yup- for all that money you spend on marketing, most of your business probably comes from referrals. Kidding- kinda 😉
Word of mouth included mentions like “referral,” “colleague,” “friend,” or “partner,” as well as Facebook group referrals, Slack recommendations, online community mentions, and specific people’s names.
While word of mouth is a byproduct of happy customers, it’s also a byproduct of a successful brand- though probably to a lesser degree. For example, many people make recommendations for products they’ve never used before, and a portion of these WOM mentions likely fit within that group.
What does WOM ranking third mean for the state of B2B?
People ask for referrals for two reasons: to develop or narrow their consideration set, or to validate their consideration set.
That means a portion of these responses (create/narrow consideration set) likely didn’t know a brand existed before they moved in-market, but another portion likely did but needed more confidence in their choice.
Either way, both situations may suggest brands need to invest more in brand awareness, association, and trust in advance of purchase.
Further research is needed to say with confidence.
Generic responses go to…
“Online” ranked fourth with 684 responses.
Online includes responses like “internet,” “web,” “net,” “online,” and “found you online.” If any of those responses included the word “search” within them, as in “internet search,” we categorized it under “search,” not “online.”
Of the data pulled, none of the forms had an SRA dropdown that included one of those terms as an option, yet many people still responded with the single, generic answer- which could mean anything, really.
Given the dominance of search and social in the data, I think it’s safe to assume that when people say “internet,” “online” or “web,” a good portion of them originated through search or social as well.
For example, If we extrapolate search (45% of responses) and social (18% of responses) over the 684 “online” responses, we can estimate that 307 came from search, 123 came from social, and the remaining 253 came from a combination of content, review sites, communities, Reddit, and display ads.
These generic mentions could also serve as a filler response from people who can’t remember how they discovered you but knew it was somewhere online.
Bottom line: We purged roughly 2,000 invalid or unusable SRA responses before running the analysis. When you combine those invalid responses with these generic responses, roughly 20% of all SRA responses were unusable.
What about blog content?
In fifth with 99 SRA responses (1% of total responses) ranks “content.”
But what kind of content?
79 people literally just said “content.” And no, none of the websites that we pulled data from had a dropdown “how did you hear about us” field with “content” as an option.
19 responses were mentions of either an article or a blog post. Whether those articles or blog posts were discovered on the company’s website or another website we don’t know.
1 response mentioned “skimming video content.”
Surprising?
I think so, especially considering that almost every single website we pulled data from had an active blog.
This could mean one (or both) of two things: first, though people have active blogs, they don’t have active distribution. Or second, the blog posts they do produce create very little memory or recall.
Lesson?
Without distribution (or ranking) blog posts will struggle to drive meaningful results. Without distinctiveness and association building, they’ll struggle to generate brand recall in the first place. Blogs posts need to do both.
What happened to online communities?
It’s hard to say definitively whether or not responses like “word of mouth” or “friends” could have also come from Slack or online communities, but only three out of 8,528 responses explicitly mentioned Slack, only six explicitly mentioned a community, and only four explicitly mentioned Facebook groups.
Note: all of those responses got grouped into the “word of mouth” category in this report.
What does this say about the emergence of dark social community referrals?
They certainly exist- we can attest to that. But further research is needed to definitely differentiate the sources of word of mouth referrals- many of which could have originated from communities.
Either way, communities are much more abundant in marketing and sales than they are in other industries, which could be the reason we saw so few mentions in this report.
Most surprising: Almost nobody mentions “podcast”
Perhaps most surprising is that only 12 people of the 8,528 responses (0.14%) mentioned a podcast in their self-reported attribution.
Was it because none of the participating companies had podcasts? No, actually- 46% of the companies we pulled SRA data from had podcasts, which suggests poor podcast distribution across the board.
Could some podcast first touches have gotten self-reported through other channels instead? After all, many companies create podcasts to splice up and distribute across LinkedIn and Youtube.
Certainly. Even still, had the actual podcast been a meaningful source of awareness and trust building, you would have expected to see more explicit podcast mentions.
Display traffic is garbage
More than half of the nearly 100 responses that came from display traffic as the last touch were incomprehensible, irrelevant, or invalid.
Responses like these were very common:
- Hi
- Yas me hear
- Jobs
- Jdehhe
- A love you
- Hxxht
- 123456
Reasons? Likely poor quality traffic, untargeted audiences, bot traffic, etc.
While display may bring volume, it doesn’t appear to bring quality- at least not in this dataset.
Past customers pay again- keep tallies on them
Coming in #11 overall with 0.74% of total responses (64) was previous customers.
Responses ranged from “used you at a previous company” to “brought you on as I was leaving.”
Takeaway? Second order revenue is real.
Happy customers leave their jobs and hire you again at their new jobs. When they leave, put them on your outbound account targeting list.
Only three people didn’t remember?
Of the 8,528 responses, only eight people mentioned that they couldn’t remember how they found out about the company.
Does that mean only eight people didn’t remember? Hardly.
Some people could have misremembered, and others who didn’t remember may not have responded at all.
In fact, we had hundreds of invalid or blank responses ranging from totally blank to “sdsiisdfg” to “…”. This underscores another challenge with SRA: people either don’t fill it out, don’t remember, or fill in something arbitrary and random.
ChatGPT, for real?
Hardly a dent in the total number of responses, but how many of you foresaw a future where prospective buyers would discover you through ChatGPT?
As more and more people make ChatGPT a normal part of their everyday workflow, expect this number to rise over time.
How can you optimize your business for ChatGPT results?
I have no idea, but I’m all ears if someone knows.
Then there’s Reddit...
Reddit came in #16th place overall with 33 self-reported attribution mentions.
So why am I talking about it?
Mainly to point out that we did not group Reddit in with social media. Reddit is an online forum inherently built around communities (subreddits), not people, and has a high degree of anonymity. So we categorized it by itself.
Reddit has a larger presence for certain industries over others, so my assumption is that Reddit might be a wild card for your business depending on your category.
And last, overlap with ‘last touch’
What percentage of SRA responses mentioned the same channel or touchpoint as last touch attribution in HockeyStack?
- Display: 2.5% overlap (2/80)
- Email: 21% overlap (20/97)
- Social: 60% overlap (779/1,287)
- Search: 53% overlap (2,646/4,939)
It makes sense that very few prospects who came through display ads or direct traffic as their last touch mentioned display or direct as their self-reported attribution. Display is largely a retargeting play targeted toward people who already know you, and direct, by definition, means they discovered you first elsewhere.
Could some of that direct traffic have come from communities like Slack that may strip referral data? Potentially. But only one person who came through direct as their last touch source mentioned “Slack” as their SRA.
It also makes sense that a larger portion of last touch search and social would overlap with self-reported attribution.
That’s because many of these journeys could have been the first touch (how they heard about you) and last touch (how they came to you today) in the same journey. Or they could have discovered you through one Facebook ad, for example, and returned via another Facebook ad later.
Either way, the majority of prospects’ last touch and self-reported attribution did not overlap, underscoring the importance of analyzing the entire customer journey, not just one memorable touchpoint.
Here’s the top 18 self-reported attribution responses by direct (last touch):
So how does self-reported attribution stack up?
One thing is clear (and if you’ve collected self-reported attribution, you know what I’m talking about): tons of responses have a big question mark next to them.
For example, when someone mentions “Google,” did they click on a Google ad, a branded query, a blog post, or something else? Or if they mentioned “article,” are they referring to a blog post on your website or an article they read elsewhere?
Self-reported attribution alone is not good at delineating the fine print.
Two solutions:
First, this is where software-based attribution like HockeyStack can bridge the information gap- at least for trackable scenarios like website touchpoints from Google. In HockeyStack, you can drill down into multi-touch sources and compare them with self-reported attribution responses for more granularity and precision.
Second, take your form field more seriously.
If you treat “how did you hear about us” (HDYHAU) like an obligatory add-on at the end of the form, perhaps people are treating it like one.
Instead, try a multi-step form that progressively asks for more details. For example, after you ask how they heard about you, ask them how they got here today and if that was different than the way they first discovered you. Sound measurement is sound experimentation.
And last, if nothing else, let this report be a reminder that no single measurement method is 100% accurate, infallible, or definitive. They all have blind spots. And that’s ok.
Sound measurement starts by combining methods into one, unified measurement stack without blind spots. That means software attribution, self-reported attribution (surveying), incrementality testing and analysis, experimentation, and statistical modeling.
Self-reported attribution is neither correlative or causal. While you can glean directional insights from SRA data in real-time, it’s not a sufficient data point for serious decision-making.
That doesn’t mean don’t ask people how they heard about you; it means just asking people how they heard about you is insufficient alone.
Think of self-reported attribution as the most remembered touchpoint of the person who books the demo.
Important, no doubt.
But what about the rest of their journey? What about everyone else in the buying committee? What if it’s really their most misremembered touchpoint? And what about the incremental lift those touchpoints have (or don’t have) on conversions?
SRA without software attribution and incremental lift analysis can’t give you those answers- and you need them to make sound decisions about your investments.
