Some people choose to ignore this tension, and repeat the old line that ‘half their marketing works, they just don’t know which half,’ while others dig deep on statistics that give them comfort on the surface, but provide false precision and create false confidence.
This is particularly present in companies where things are actually going well on the surface; growth is good, sales are getting results. But there’s also that feeling at the back of your mind, that the results are not as repeatable or as predictable as you would like. Marketing's impact is unclear, and maybe there’s a lot of reliance on a few individuals.
New software tools using AI promise to fix some of these problems. In reality, bolting these on to the complex machinery of the average B2B pipeline is just as likely to magnify existing problems and create new ones, as it is to solve them.
This is the second-order problem most companies miss. AI won’t be able to scale up, or fix problems in, a marketing setup that it can’t read. The first step to fixing your marketing is to make your system legible. Meaning it becomes something that both people – and AI – can understand.
The pipeline, no i mean the whole pipeline
Most companies define pipeline narrowly, looking at the stages of their active deals. But the real pipeline both starts earlier and ends later, and that gap is where most marketing waste lives.
To really understand what is happening, the entire journey of your customer needs to be tracked and seen as part of one single system. From your perception and position in the market place, to the first time someone hears about your company, to their first contact with you, and then the long-strange journey they follow to eventually become a customer.
Without this full picture, marketing campaigns are run in isolation. They end up driven by tactics, platforms or random ideas, without really considering how and where they will impact the pipeline.
Marketing and sales operate on different realities
A common situation in Australian B2B companies is an almost complete disconnect between the way marketing and sales view their roles, the pipeline and the language that they use. They each work with their own map, but neither map is the full territory.
Companies in this situation usually have well-established sales teams who are crucial in driving revenue, whereas marketing is more likely a smaller, newer team. Sometimes just a single marketing co-ordinator, or no internal marketing function, with decisions made by others, and work outsourced to an agency.
In these scenarios, marketing’s map revolves around campaigns, platforms, copy and collateral, and if their performance is measured it’s based on statistics; web traffic, social post impressions, and sometimes leads generated.
Sales meanwhile are focussed on relationships. Maintaining contact with long-term leads and past customers, waiting for them to be ready to buy again, and chasing long-term deals, providing the necessary info, clarifying and quoting based on clients needs, and chasing down responses when so many of these prospects become unresponsive.
There is a limited connection between the work the marketing and sales teams do, and no consistent feedback loop between them. In some cases, this disconnect causes active mistrust; marketing feel sales don’t value their work or follow up their leads properly, sales are dismissive of the poor quality leads that marketing feeds through, and does not see them as core to their actual real pipeline.
Attribution is either missing or misleading
If attribution is tracked at all, it starts and often ends on the marketing side. Meaning that marketing tracks statistics to measure the ROI of their work, but these don’t link up with what the sales team do.
This means that there is no attribution of where the actual closed customers came from. When statistics are used, they tend towards the easiest things to measure (traffic, impressions etc), over-weighting their contribution.
There is usually a strong first and last touch bias (or even just a last-touch bias), i.e. focusing only on the last trackable interaction before a customer closes; such as Googling you, using your contact form etc – This data is usually the easiest to measure, but that does not mean it is the most useful or the whole picture.
The first and last touch only approach underestimates the myriad other interactions that most likely took place.
The customer journey is non-linear
The reality of a typical B2B customer interaction is long, convoluted and sometimes circular. To begin with, customers might be aware of you through your long-term marketing presence, or through a referral, or an employee who previously worked for a former customer.
When they get in touch, they might search your company name, or enquire via your website or engage with a social channel; but often that could be because they don’t have the right email or direct line handy when they want to get in touch. That platform alone is not the first thing that made them want to contact you, though it may be all that is tracked.
Along the way, multiple people may be involved in the decision – and quite often the actual people at the prospect company will change. Your sales team might meet new contacts at a trade show, but also reinforce and move along the conversation with existing contacts, both on the expo floor and in the evening hours.
After the trade show, your contact might discuss your offering with their team, and one of their colleagues will look for more info, perhaps they’ll request information, or just read your website, or dig out a newsletter you sent six months ago and click a link in there.
The point is it’s messy, it’s non-linear and involves not just multiple touch points, but multiple people as well.
Many of these interactions are not visible in the typical B2B reporting system. At best, the sales team might record each interaction properly in the CRM, but key details remain offline, in their emails, phones, or just in their heads.
Marketing might track the online interactions, but won’t be able to apply them to the right account or deal in a useful way.
Automation and AI
All of this adds up to a system where nobody has complete visibility or understanding. Apart from the missed opportunities and wasted spend, the real danger comes as pressure increases to add new software, automations and AI.
This could be adopting a CRM for the first time, or upgrading an existing platform, or bringing in marketing automation, or now adding on new tools that promise to use AI to better track attribution and ROI. Starting with the software, and not the underlying system, risks creating something that at best wastes a lot of money and effort, and at worst amplifies existing problems.
This is where second-order thinking matters. AI applications like Claude Cowork can now pull all the info out of platforms like hubspot and create great looking reports whenever you need them. Whether you have all your data in one software platform, or more than one, you can link them up, end silos and create live dashboards and reports.
But if knowledge lives in salespeople’s heads, desktop drawers, and relationships; AI can’t see it. AI rewards companies with explicit, structured systems – it only works when the system is legible. If you plug it into a system already built on incomplete data and assumptions, it just makes this wrong information sound even more convincing.
AI solutions for attribution can work well, and detect patterns that humans miss, but only if they are working with large enough data sets, and more importantly, the actual data. If it is working from marketing stats around traffic and impressions, and a very incomplete picture of sales activity, AI will draw the wrong conclusions from the wrong information, missing the actual interactions that made the difference to the deal.
This can end up giving you even more misguided false confidence in things that are easy to measure, but possibly having little to no impact.
How on earth can we fix all of this?
The good news is, all of this is solvable. As AI and other tools improve, it’s easier to fix. It can be a big undertaking, but it can also be broken down into discrete areas, all of which will have some immediate benefit.
The first and most important step is to map and understand how deals actually happen. While a lot of work starts at the marketing end, if you are new to this, it is much better to start at the sales end.
Conduct a few ‘Opportunity Journey Reconstructions’ – Look at your actual closed customers, and track their journey backwards to understand where they came from, and how they got there. It’s also helpful to look at deals that were lost, and map their journey backwards as well.
Often this shows up the marketing/sales disconnect – marketing generates leads and sends them into the ether, and sales pulls deals out from elsewhere in that ether.
The goal is to uncover those missing stages and line up the pipeline in the middle, then you can get a clearer picture of where the marketing activity has led, where deals were lost, and where the best customers have come from.
Next, you need to align marketing and sales around this reality.
For marketing, this can mean abandoning some cherished beliefs and customary activities. For sales, it can mean taking more responsibility to clearly record interactions.
Typical examples of this in practice could include;
- Abandoning or winding down marketing channels that contribute little; e.g. poorly managed SEO or SEM campaigns that are driving low-value, unqualified web queries that don’t result in business.
- Expanding channels that do work, such as those which coincide with positive decisions.
- Creating campaigns that target weak points in the pipeline, e.g. if the sales team report that they have the support and intention of technical buyers, but end up losing the deal when management or finance get involved.
- Creating materials and content that answer questions and overcome common objections that sales come across.
- Building user-friendly, low-admin systems within your CRM that make it easier for the sales team to track key information on where deals do or don’t progress.
Accept ambiguity
In this complex environment, not everything is easily trackable or visible, so while on the one hand you want to get as much clarity and visibility as you can, it is also important to accept a degree of ambiguity. Not every deal will be able to be mapped perfectly. Not every valuable marketing channel can be clearly reduced to statistics. The long-term efforts to build your brand play a bigger role than the last tactical campaign, but are a lot harder to quantify.
This means to some extent you end up using a combination of anecdotes, common sense and borrowed wisdom from larger, analogous companies, alongside the best data you can get out of your own activities. Accepting a little ambiguity here is not a sign of failure, it’s about choosing the right level of resolution, to get enough clarity to make better decisions than you could yesterday.
Once you’re measuring the right things, and have a working map of how your complete pipeline is working, you can continue to iterate and improve. You can test an assumption, and if the results still don’t line up with your expectations, you’re a lot closer to understanding why, and can tweak the system accordingly.
This is where the benefits of automation start to really make sense.
Most of the issues with marketing in B2B companies comes back to not having visibility of the system, because it is illegible. The information exists – but it’s scattered across people, conversations, and systems in a way that no one can see end-to-end.
Until you can see how revenue actually happens, you can’t improve it. The first step is to map your pipeline properly, confront the uncomfortable reality to find out what you don’t know and can’t see, even when things are going well on the surface.
