What ice cream sales, AI, and sunburn tell us about pharma marketing

If you ask BARD (Google’s AI equivalent to Chat GPT) a deliberately silly and loaded question such as:

“If sunburn rates are higher when more ice creams are sold, shouldn’t there be warnings about the dangers of eating ice creams?”

BARD is smart enough to highlight that although there is a correlation between ice cream sales and sunburn rates, it does not mean that ice cream causes sunburn. It highlights something called ‘weather’ as being a common factor in both, pointing out that correlation does not equal causation.

However, if you ask BARD a much subtler and less silly, but very similar question, such as:

“If field team calls that include digital content drive more patient starts, shouldn’t we use more content in all sales calls?

BARD simply says yes, we should use more content in all calls, and then gives you a bunch of tips on using more digital content. But is this another example of correlation not equalling causation, or does adding more content really ensure all calls are more effective?

Intelligence beyond AI

Interestingly Veeva, who has a tonne of data on this stuff, suggests the same as BARD – that content use in field team calls equals greater effectiveness. In fact, they recently highlighted that field team calls that use digital content, drive 2.5 times more new patient starts than calls without content – which certainly seems to back up BARD’s answer.

However, Veeva knows a couple of additional things BARD probably doesn’t, which is that:

• 77% of pharma content is rarely or never used by field teams
• 61% of field team calls contain no digital content.

This to me is a bit odd. How can using content drive significantly better calls, and yet, more often than not, field teams don’t use any content? To me, A bit like the weather in the sunburn example, something critical is missing from this equation.

What’s really going on

Deciding to use a new or different medicine typically takes time – it can literally be a life and death decision and it’s typically complicated. It’s a journey rather than an instant choice – a journey we sometimes call the adoption ladder, and it can take months and many, many touch points to move from not prescribing at all, to prescribing for the first time. This means the measurable effectiveness of an isolated call in this journey is hugely impacted by where that doctor is on the adoption ladder at that moment, for example:

• If the prescriber is completely unaware of the medicine, then the chances are that a single sales call with the field team is going to be a ‘pop-in’ moment with limited attributable effect
• If the prescriber is already actively considering prescribing the medicine for the first time, and they ask for a specific appointment, the field team member might have 20 minutes or more to present data and answer questions – and the likelihood of an attributable impact is really very high in this scenario. ?

In isolation, one of these calls, the one that happens to have a lot of content, looks like it correlates with a big increase in patients starting treatment. However, it’s not more effective because it was longer or because a lot more content was included. It was longer and more content could be included because the prescriber was interested in prescribing already and explicitly wanted more information.

Measuring the impact of individual calls like this, is like giving a pay-per-click search ad 100% of the credit for a sale when a buyer searches for the brand name. The fact that the person searched for the brand name tells you they were interested already. And, most likely a bunch of other touch points created that interest – but because it’s simple to measure we give all the credit for the sale to the search ad.

In a lot of ways, it’s the same as the ice cream example – the correlation between content and patients starting treatment does not equal causation. What really caused the sale could be all sorts of other things that might be very hard to track.

What we can’t learn from ice cream sales, AI, and sunburn

The whole point of this newsletter is to present something other than the standard industry narrative. In part, that’s because BARD or Chat GPT or any other AI tool can now chop together a bunch of conventional wisdom at the click of a button. However, commercial advantage doesn’t come from doing what everyone else does or from using BARD – it comes from achieving a different level of understanding that goes beyond the norm.

Nobody is specifically wrong here, but I think they are missing something valuable. Lots of field team calls don’t use content because lots of call settings don’t really allow it. That doesn’t make those call settings ineffective – they are often critical to generating high-interest calls. However, really getting to the subtleties of what happens in each of these calls can help improve the content we create. And that in turn will improve our overall effectiveness.

For example, Veeva also highlights that four slides is what is typically covered in a call with content – has anyone ever designed a visual aid with a four-slide flow option? I doubt it. And yet it would surely be helpful given these data, wouldn’t it?

When we really understand what people are doing and what motivates them to do it, we find an edge – an advantage. This comes from seeing the whole journey not just the bits the hard data captures. AI can be helpful, but it can only work with the data it has access to – and if that data is the same gap-filled data everyone else has, it’s not helping us win.

So, what can we learn from ice cream sales, AI and sunburn? About as much as we can learn from most of the industry narratives out there – nothing different, nothing that gives us a real advantage. For that, we have to get closer to the people in the equation – the field teams and the prescribers. The better we really understand them, the details of their interactions at a human level, and all the hard-to-measure stuff that sits around them, the more advantage we’ll find.