My conversation with Andrew Grant on why smaller companies still need third-party media

For the past month, I've been on a mission to make scientific content great again.

Alongside surveying 100+ scientists, that's meant reaching out to marketers, most I've never met before, to talk about how we can actually do better for the people we're supposed to be writing for. 

Reaching out on LinkedIn these days is hard. Everyone's immune (rightly) to asks that might secretly be a sales pitch in disguise, so it's always nice to get a real response back and jump on a video call. Turns out LinkedIn is still useful for something, as long as you avoid the pitch-slapping.

My conversation with Andrew Grant last week was a great one. No stiff small talk before we got to the good stuff, just two people talking shop. The hard data's being saved for the industry report that's coming. This is the conversation underneath it. Let's get into it.

Andrew is Digital Marketing Manager at SPT Labtech, running digital across their full device and product range. 

A few things we covered:

• The time an AI-detection tool told him his AI-assisted content was written by a human

• Why the scientists everyone wants as named experts are usually the ones least able to give it

• A genuinely different take on third-party media sites to the one I heard last week

• Why his team walked back gating, then partially walked that back too

We started on how much AI-generated content has actually improved. Andrew commented that AI content across the industry used to be thin, easy to spot, not saying much. That's changed enough that he's had a genuine surprise recently.

"I tested a web page I'd built with AI to one of these tools that scrapes it to check if it's AI generated," he told me. "It came back and said no, this is human-written. And you're like, well, it isn't." 

His theory is that the more specific the science and chemistry get, the harder it is for detection tools to flag it, the content simply doesn't read like the generic stuff they're trained to catch. Now this content is still reviewed and edited by the team within the business, but it proves that these tools are improving exponentially over time.

That doesn't mean he trusts the label either way. He made a similar point to one I heard last week: people react to content differently depending on whether they know AI made it, regardless of the actual quality. 

"If you just put the fact aside that this has been generated by AI, you'd go, yeah, this is really good," he said. "But because you know it's AI, your instinct is to go, I don't like that."

I think we’ve all felt that with obviously AI-generated outreach emails that don’t sit well, not because the content's wrong, but because nobody could be bothered to write you a proper one. Andrew put it more bluntly than I did: "it leaves a sad taste in your mouth immediately."

We talked about named-expert content next, the thing every life science marketer in this series agrees is the strongest trust signal available. Andrew put the problem better than I have.

"They want it, but they won't give it. That's the challenge," he said. 

The scientists whose validation would carry the most weight are often the ones tied up at a large pharma company, working on something they're contractually not allowed to discuss. The people most qualified to be the named expert are frequently the least available to actually be one, often not because they don’t want to, but because they can’t.

I pushed him on whether that could work internally instead, a piece with SPT's own head of R&D attached rather than an external voice. His answer was honest rather than convenient: an external key opinion leader is always going to be the strongest source, because they've got nothing to gain by praising the product. An internal expert, named and credited, still beats an anonymous corporate blog by some distance, but content built purely around an internal voice, even a genuinely senior one, tends to get read as sales material by a scientific audience, whether or not that's the intent. It's not the best option, but still a solid second.

On third-party media, Andrew's take was a genuinely different angle to what i've heard recently:

"We're not one of your traditional big players in the market," he said. "How do we get into that kind of conversation where we're being referenced all the time? We have to go to third parties to try and reach people, because otherwise, if we just push stuff out ourselves, it's not going to get the reach we want."

He was honest that results are "a mixed bag," and that he tries not to judge every placement the same way. "A webinar is not going to have the same KPIs as an eblast, or the same response in terms of opportunities generated," he said. Some of it isn't bought for leads at all. "It's going to generate brand awareness, and as much as the individuals within the business might ask, ‘well, what does brand awareness do?’, we know as marketers that we've got to be out there so people know who we are." 

Different size of company, different demands.

Gating followed a similar pattern of trial and partial reversal. SPT moved away from gating most content a couple of years back, after deciding they'd rather people actually engage with it than sit behind a form collecting names for an internal leads report. These days it's more selective: some content gets gated for a month after release, then opened up entirely once that initial window has passed.

I've landed in a similar place with clients recently, for the same reason. If the content is genuinely good enough to be worth reading, gating can be the wrong call. Leave it open, let people actually see it, and trust that it brings leads into the funnel some other way - based on the value it provides. Where I do still recommend gating, it's usually a second piece, something linked from the open content once the audience has already seen enough to trust the quality and is willing to hand over their details for more of it. 

Gate the follow-up, not the proof.

His closing point was the one I keep coming back to since the call. AI has let his small team achieve more than they could otherwise manage, but he was clear that speed isn't the point if it costs quality. His test for anything before it goes out: would he be happy if he landed on someone else's site and saw the same thing. If the honest answer is that it looks not up to scratch, it doesn't go up, however much time it saved getting there.

Thanks for your time, Andrew Grant !