We have spoken previously about the impact of ‘dark social’ for B2B marketers and I was interested to see the topic hit the headlines again this week with research carried out my Rand Fishkin and SparkToro. I’ve followed Rand since his days founding SEOMoz, which later rebranded into Moz, so when this research came out I knew it would be meaningful and significant.

The concept of dark social is not new, it actually dates back to 2012 and a piece in The Atlantic, but what is newer is the amount of attention it is getting in the B2B world. To put it starkly, its impact is enormous and most of what you think you know based on your web analytics is likely wrong.

What is dark social and how does it relate to B2B?

Dark social, otherwise known in B2B communities as ‘dark funnel’, ‘dark traffic’ or ‘dark social media’ – I actually think ‘dark traffic captures the term best and I’ll show you why below –  is a concept used by marketers to describe the web traffic that comes from popular modern distribution channels which is incredibly difficult to accurately track. These are places where B2B buyers are highly active, and are places which are directly impacting your funnel but you will not (currently) have direct visibility of the impact.

When we say ‘modern distribution platforms’ we can break these down further into;

1. Employee communication platforms; think Slack, Teams, Zoom and any other internal comms tools that you may use
2. B2B groups and communities: Industry Slack communities, Clubhouse, WhatsApp groups, Facebook Groups, LinkedIn communities and anywhere that can groups of likeminded professionals can gather to share information privately
3. Word of mouth and in-person recommendations: formal gatherings like industry events all the way through to informal gatherings, smaller pub gatherings of industry friends etc
4. Content platforms: podcasting platforms like Spotify, Apple podcasts, YouTube
5. And of course, Social networks: LinkedIn, Quora, Twitter, Meta platforms etc. Social can br broken further into private community based social networks like Discord and the more ‘open social’ platforms like Facebook.

As you can see above, ‘dark social’ is really only part of the attribution challenge, we’re really looking at a dark traffic problem.

Increasingly the go-to social networks for discussion are also increasingly the private ones. This makes tracking them increasingly difficult. Where once, social focussed on being open and broadcasting publicly, slowly but surely the major social networks brought in more privacy tools, from Facebook down to the likes of Twitter, and then more recently B2B decision makers migrated into more private forums, including the likes of Discord, Slack channels, etc. Discussions which may once have ben somewhat open in Twitter threads are now increasingly hidden in WhatsApp and Slack channels, making tracking them neigh on impossible.

To make matters worse, 84% of B2B web pages that get shared across both vendor/client pages and publisher pages is now done in private channels such as Slack, email and instant message. To me this makes perfect sense, I don’t find an interesting tool, service or piece of content and rush to Twitter or LinkedIn to post it, I am, by an order of magnitude, more likely to Slack it to the relevant team members within FunnelFuel or text to the relevant stakeholder. To compound it further, these are the EXACT sort of shares which are most likely to lead to  B2B purchases.

So all of this paints a fairly bleak picture when it comes to measurement, tracking and measuring the business impact with your web analytics platform.

Google Analytics and Dark Social – now you know why you (apparently) have so much ‘direct traffic’

Accurate information is an expectation of site owners and marketers when they check their analytics tools to see how visitors find and access their site. Sadly, this data is prone to being seriously flawed. Recently, SparkToro collaborated with Really Good Data in an experiment which entailed driving 1000 visits from 11 popular social networks. The aim was to study how Google Analytics categorised these referrals. The experiment was robust, and they pulled in an expert resource from GA4 migrations and implementations, a vendor which specialises in setting up GA correctly.

The results were stark. A number of major social networks obfuscate their referred traffic entirely by stripping out any refer information, and a significant, additional set obscure referral data at least some of the time. That second set seems almost intentionally misleading; technically you’d expect to see refer either populated or not, so removing it occasionally speaks of darker arts.

As per SparkTuro – this was their findings: Article: https://sparktoro.com/blog/new-research-dark-social-falsely-attributes-significant-percentages-of-web-traffic-as-direct/

• 100% of all visits from TikTok, Slack, Discord, Mastodon, and WhatsApp were marked as “direct,” and contained no other referral information.
• 75% of visits from Facebook Messenger contain no referral information. This does not appear strictly related to browser choice, device type, or web vs. app.
• Instagram messages (DMs) as well as public LinkedIn and Pinterest posts also missed substantial portions of referral data (30%, 14%, and 12% respectively)
• A smaller amount of traffic was misattributed to “direct” by Reddit posts, LinkedIn messages (DMs), and Twitter DMs
• YouTube, public Instagram profile links, public Facebook posts, and Tweets appear (for now) to provide referral data in most or all cases

I have marked in bold those that have the biggest impact on B2B purchase decisions, although a case can be made for all of these channels having an impact when you consider the full width and breadth of B2B services and offerings, especially and most pertinently for any brands operating in B2B2C.

All this means that if you take GA at face value, then you are misrepresenting everything from lead to sale against ‘direct’ when in reality you can expect that dark social channels are actually a significant portion of those refers.

What can B2B marketers do to build a better picture of what is actually happening on their website and how can we start to uncover the true origins of our B2B web traffic?

• Firstly, if you are a B2B marketer who bases a lot of your assumptions on your users and how they found their way to your website on GA, then we recommend you revisit that assumption
• This dark social traffic has no or limited refer information appended, when GA sees that, it is categorising the traffic as ‘direct’. It would be better to re-categorise that internally to reflect the true broader array of traffic sources
• Dark social channels like Slack communities, Discords and WhatsApp realistically ARE impacting your sales funnel but they’re not showing up in GA, meaning they quite possibly are not been considered at all by many organisations. This means that any strategies that could be deployed to make these channels work harder are being missed alongside the more obvious reporting metrics mis-matches.

How can FunnelFuel’s Journey platform help you get more precise web traffic data?

1. FunnelFuel tracks 100% of all web visits to your brand or B2B publisher pages. GA and other analytics platforms use sampling and predictive analytics to ‘guess’ and approximate. This is important because this gives us a better chance of catching the real refer information when it is present – so that closes some of the data loop.
2. Utilise our campaign URL generator: This provides a unique URL formatting service which makes it possible to append your own refer data to the traffic coming out of social media and other such places. It doesn’t entirely solve the problem but it does fill in a lot of blanks. Getting into the habit of using this tool across key social distribution channels will give you much more robust data to work with within FunnelFuel.
3. We integrated the browsers of major social networks into our tracking: The likes of LinkedIn and Meta have their own web browsers that they open outbound clicks into – by integrating LinkedIn and other key web browsers we are able to (realistically) infer that a ‘direct’ click is really a LinkedIn click.
4. We provide real-time (live) log level data which captures much more ‘raw data’ and sometimes this captures more refer data or other signals that can help draw deeper inferences. An example of what this can pick up is server re-directions which may leave a trace on the true origins of the click
5. Track deeper page lands – this is a method of inferring data. It is worth pointing out that true direct traffic is much more likely to land on your homepage and is not likely to land on some deep URL. Somebody could type in funnelfuel.io based on hearing about our services but they are not going to type in “https://funnelfuel.io/scaling-b2b-seo-ai-content-programmatic-seo-tactics-to-dominate-search-engines-lower-sem-costs/” – so realistically and ‘direct’ entries into that page are going to be link clicks not type ins.
6. Ultimately – we have to accept that when a referring platform has either not provided refer data or sometimes strips it out we can not generate this – if it is missing it is missing and there will be data leakage. No platform is perfect 100% of the time, especially when some social sites seem to be intentionally misleading

We’re also dedicating time to thinking through how we can uncover more of this traffic, and we’re working to solve these problems for B2B marketers.

For many the first step is understanding what is dark social and how it is impacting your pipe today, and then getting better and using unique URL generator tools, full visit tracking and other smart tactics to build a bigger picture of how these dark platforms are impacting your business.