“Danceability” is Dr Simon Hefti’s term for the missing ingredient in most sales data: the extra layer of explanation that turns a flat metric into something a salesperson can actually use.
We spoke to Simon, co-founder and chairman of D ONE, a Zurich-based data and AI consultancy, who argues that most sales organisations aren’t short of data, they’re short of the insight that makes that data usable.
Key takeaways
- Data has no value until it becomes actionable – it needs to change what someone does next, not simply describe what has already happened.
- “Danceability”, a term borrowed from a study of 50 years of Swiss music charts, describes the extra layer of explanation that turns a flat metric into a story people can act on.
- Any data-driven suggestion needs an explanation attached to it, or salespeople will ignore it.
- AI and automation work best when they free up time for what only humans can do: build relationships and read a room.
- What happens inside a sales conversation matters more than how many calls or visits a salesperson logs.
- Large data sets carry a real risk of confirmation bias, so findings need to be validated – or “back-tested” – before anyone acts on them.
- Using customer data well requires an ethical filter, or personalisation quickly tips over into something that feels intrusive.
- Sales leaders don’t need to become data scientists, but they need the ability to ask better questions of the data they already have.
When does sales data start creating value?
Collecting data is only the beginning. As Simon explains, its commercial value emerges when it helps someone make a better decision, change an action or influence an outcome. In this episode of Mastering Sales & Negotiations, he traces that value back to four basic levers: increasing revenue, decreasing cost, decreasing risk and, most ambitiously, finding entirely new sources of income, such as moving from a product sale to a subscription model.
The common thread across all four is actionability. He compares it to a weather forecast that tells you it’s going to rain. The forecast alone has no value if you don’t pick up an umbrella. Data only becomes valuable at the point someone uses it to do something differently – and, in his experience, plenty of sales organisations are still going out without an umbrella.
More data does not always mean better insight
The challenge facing many sales organisations is not the volume of information available. It is whether people trust it, understand it and know what to do with it. Simon points to a pattern he noticed while speaking to sales leaders shortly after ChatGPT’s launch: every one of them had a CRM system, and not one believed the data inside it was accurate.
He splits the issue into two parts. The first is straightforward: the accuracy, availability and recency of the data itself. The second is subtler and is the one he keeps coming back to throughout the conversation – whether the data has been given enough context to actually mean something to the person looking at it.
What can “danceability” teach us about sales data?
Simon’s idea of “danceability” offers a useful way to understand why some data informs action while other data simply fills a dashboard. A number may show what happened, but it often needs another layer of context before it can explain why.
He draws a direct line to sales data. A win rate on its own is just a chart position. The value comes from finding what drives it – a product specification, a particular go-to-market approach, or the behaviour of a specific individual. Clustering techniques can group people who behave similarly and match that against their results, so that the best performers’ habits can be identified and shared across the team. One of D ONE’s clients takes this to an extreme, tracking roughly 150,000 measures across its customer base every month to build a genuinely detailed picture of what matters to each one.
How can sales leaders challenge the story their data tells?
Finding a pattern is not the end of the analysis. Simon argues that organisations should actively test whether an apparent relationship stands up before allowing it to shape a commercial decision.
The second safeguard is explanation. Simon is direct about what happens when a salesperson is handed a recommendation with no reasoning behind it: they ignore it. “I know my clients, don’t tell me what to do” is the instinctive reaction. Pairing every data-driven suggestion with a clear explanation is, in his view, the difference between advice that gets used and advice that gets binned.
AI, intuition and the value of human judgement
The conversation is not really about choosing between technology and human instinct. Simon’s view is that each becomes more useful when organisations understand the role it should play.
Resistance often reflects how technology is introduced rather than a blanket unwillingness to change. People are more likely to engage when they can see how it will reduce administration, support their judgement and give them more time with customers.
Framing means being explicit that automation exists to free up time for judgement and relationship-building, not to replace it – removing low-value admin, such as flagging a bounced email back to the CRM automatically, rather than removing the salesperson’s role in the conversation. Upskilling means giving people hands-on exposure to what these tools can do. Simon also notes an interesting generational trigger: leaders who see their own children using far more advanced tools in newer companies often experience it as a genuine turning point in how open they are to change.
Looking beyond the rear-view mirror
Historical measures remain useful, but they cannot tell salespeople everything they need to know about an opportunity today. Simon encourages leaders to look for smaller, more immediate signals that reveal how customer interest is changing.
His second point is about resolution rather than timing. Working only in averages – average deal size, average sales cycle – is, in his words, like driving “a big sedan” that glides over the road without letting you feel the bumps. The leading indicators worth watching are the small, individual signals: a shift in a specific customer’s engagement, a change in the tone of their questions, a stretch of silence that breaks an otherwise regular pattern of contact.
What happens inside the conversation matters
Calls, meetings and emails are easy to count. The harder and more commercially useful question is whether those interactions changed the customer’s thinking or moved the opportunity forward.
This is also where the continuation versus advance comes in: whether a customer is taking genuine action to move a deal forward, or whether the salesperson is doing all the work while the customer stays passive. Tracking that distinction across a number of interactions flags problems – or opportunities to accelerate – far earlier than a monthly win-rate report ever could.
When useful insight becomes unwanted intrusion
The same data that helps a salesperson become more relevant can also make an interaction feel uncomfortable or manufactured. The difference lies in how thoughtfully the insight is interpreted and used.
Another mistake is treating automation as a purchase rather than a process. Customer-facing chatbots are a good example: useful for answering routine questions, but rarely used for their most valuable purpose – spotting the products customers are searching for that the business doesn’t yet offer. Simon’s view is blunt: if a business hands the problem to IT without first working out which questions actually matter to them, the initiative won’t deliver – the missing ingredient is time spent tailoring the tool to the business, not the tool itself.
Connecting insight with sales coaching
Data becomes more useful when someone can translate it into a relevant and achievable next step. In Simon’s example, that may be a suggested action in the CRM, supported by an explanation rather than presented as an instruction.
Interestingly, when D ONE has run A/B tests comparing teams with and without this kind of support, win rate often barely moves – but deal volume increases. Layer in real-time signals, such as a customer’s engagement rising or falling, and that same coaching relationship becomes a genuinely live feedback loop rather than a monthly review of what already happened.
Where should sales leaders begin?
A useful starting point is not another dashboard, but a closer look at the conversations already taking place. Simon encourages leaders to explore what customers are saying, what changes during an interaction and whether either party takes meaningful action afterwards.
His second piece of advice doesn’t require any technology at all: take half a day away from the day-to-day and genuinely think through what information you wish you had but don’t currently have – then start working towards the most important of those gaps. None of this, he’s clear, requires a sales leader to learn to code or build models themselves. The real skill in demand is the ability to ask good questions of the data that already exists, and to resist the natural pull towards counting only what’s easiest to count.
Turning better insight into commercial performance
The central lesson from the episode is not that sales organisations need to collect more information. It is that the information they already hold must be trusted, explained and connected to a useful next step.
As Simon puts it, we consistently underestimate the intelligence we already have. The opportunity for sales teams isn’t to hand decisions over to AI, it’s to use it to surface the patterns a human then has the time, and the story, to act on.
This episode of Mastering Sales & Negotiations is available now. Follow the full series here or wherever you listen to podcasts.
Dr. Simon Hefti is a leading voice in data-driven value creation, helping organisations turn AI and data into measurable business impact. A physicist by training and co-founder of D ONE and herlock.ai, Simon combines deep technical expertise with real-world commercial insight, making complex technologies practical, accessible and valuable.
