B2B marketers are experimenting with GenAI for content marketing tasks from personas to content marketing strategy to messaging, content, social posts, and more. Most are also doing so without a strong foundation for validating the outputs ChatGPT or its peers produce. We know that the output from basic prompts is mediocre. My concern is whether we know how to fix it. Or avoid most of it from the start. A Buyer MVP (minimum viable persona) can solve this problem.
In recent work with clients, I’ve found a Buyer MVP helps marketing teams catch and correct the flaws in meaning, context, truth, and brand representation in the outputs from LLMs. They’re able to strengthen the relevance of the outputs to create more effective content. This enables them to engage buyers without risking the downsides that come with these early-stage LLMs.
But we also need to be aware of the traps we can fall into with LLMs if we’re not paying attention. The biggest one is that we can think of ChatGPT (or one of its peers) as human. When we do this, we’re less likely to question the outputs. Afterall, an LLM is pretty convincing when it tells you things.
The bottom line is we must remain skeptical and come armed with the information we need to validate those outputs. Not just for the sake of buyer relevance, but for brand relevance. An MVP is one tool that helps keep you grounded.
Why We Must Quit Ascribing Humanness to LLMs
I see articles, posts, and interviews where marketers speak about their LLM of choice as if it’s a human. I see prompts where marketers add please and thank you – as if they’re having human, empathetic conversations with the LLM. It’s a machine. Yes, it’s programmed to try to please you, but it’s still a machine without empathy, context, or a worldview.
Last week Lattice announced they were onboarding AI agents as employees – then quickly reversed that announcement given online backlash. A statement from their CEO said the company has stopped its pursuit of digital workers in the product. Humans found it insulting.
Yet we still succumb to GenAI’s seeming humanness. “As humans we have this ‘weakness’ that if something communicates with us in natural language [we think of it as if] it is a human,” says Aleksander Madry, a professor on sabbatical from MIT, leading a team at OpenAI called Preparedness.
Yet, LLMs are not human. An article in Scientific American explains that we perpetuate the anthropomorphizing of GenAI by categorizing the mistakes and inaccuracies it makes when stringing words together by calling them hallucinations – a term related to human psychology. Instead, the authors say, we should label such outputs as bullshit.
“When someone bullshits, they’re not telling the truth, but they’re also not really lying. What characterizes the bullshitter is that they just don’t care whether what they say is true. ChatGPT and its peers cannot care, and they are instead, in a technical sense, bullshit machines.”
Let’s flesh out what “bullshit” really means in relation to your LLM of choice.
Consider that GenAI:
- Doesn’t know the difference between truth and falsehoods.
- Can’t reason to guarantee its inferences are correct.
- Is incapable of fact-checking its own work.
- Has no empathy or contextual understanding of what you ask.
- Gives you statistically probable responses (disconnected from meaning and fact) based on your prompts.
Given all the above, is a bullshit machine what you want to put in charge of how you represent your brand and engage with your buyers?
Why B2B Marketers Need a Buyer MVP (Minimum Viable Persona)
It’s reasons like the above that B2B marketers must be smart about how they engage with GenAI. Efficiency is not worth much if you sacrifice effectiveness in the process. There’s a lot of advice to upload your personas to an LLM to help steer them in the right direction. The problem is that many personas are built on mediocre templates, maybe using some data, but without buyer/customer insights. Their input is not much better than GenAI’s output.
This may be why a Forrester study, Customer-Centered Messaging Motivates Buyers to Act, found only about 1 in 4 marketers speaks the language of their audiences and decision-makers. It’s hard to create content that resonates when your buyers must expend tons of effort to grasp why it’s relevant to them. Most times they won’t bother.
An MVP gets you past this obstacle with less effort than building a full-blown buyer persona.
Benefits of a Buyer MVP (Minimum Viable Persona)
- Less Time: Marketers push back on buyer persona development due to the effort and resources needed to produce them. You can build an MVP with as little as 5 – 7 customer interviews. Think a few weeks vs. a few months.
- Baseline for context: Understand why buyers need what you’re selling. Know why they care about solving the problem, and how they frame the problem against their objective.
- Common attributes: Learn what people representative of a buyer persona have in common, including the words they use. This helps you engage a wider swath of a specific target segment.
- Impacts to momentum: Grasp the emotions in play professionally. Uncover questions they must answer to move forward, and what concerns and obstacles slowed them down.
- Preferences for engagement: Understand which channels they use (and for what). Gauge the voice and tone and format preferences that help them engage.
With these baseline insights, you’re now primed to employ the assistance of an LLM to flesh out the bones of your MVP. And use it to produce relevant storylines for engagement programs, content ideation, writing assistance, and editorial reviews.
Just how does that help, you may ask…
As Heather Murray says in her recent newsletter on writing killer prompts:
“Detail is critical, and it’s what will differentiate your results from everyone else’s too. I always say to my trainees: ‘if someone else could use your prompt, it’s not specific enough to you.’”
I’ve built hundreds of buyer personas over the last 17+ years. No two of them have ever been the same. Whether it’s perspective on the problem to solve, the industry they’re in, the attributes of the ICP, or other factors at play.
Specificity is important. Context is critical. Relevance is non-negotiable.
If you don’t know what you don’t know about your buyers, your chances of getting strong outputs from GenAI are slim.
The Trouble with Data-Only Buyer Personas
And one more thing. If you think can do all the above by using data as the basis for your personas, think again.
Access to data changes the game for B2B marketers. But it also has its limitations when compared to the insights from customer and buyer interviews. Conversations are necessary for building in-depth Buyer MVPs that inform profitable content marketing strategy.
For example:
| Data Tells You | Data Can’t Tell You |
| What someone did (content viewed) | Why they did it |
| How long they spent doing it (dwell time) | If they got what they wanted |
| Where they came from (source) | How they explain the problem they’re solving |
| How often they visit (frequency) | If they found the experience valuable |
| What they searched for (keywords) | What they would’ve liked to see or learn |
| What words and phrases they normally use |
And some of what data can tell you is only applicable if you know WHO took the actions. If you’re marketing to Directors of IT but consultants are the people taking the actions, are the insights from that data relevant?
The problem with data is that it doesn’t paint a full picture. Therefore, marketers are making guesses based on wishful thinking unless that content is in a controlled environment where you know who’s interacting with it. Nurture streams deploying content hosted on blind pages with access via links in emails sent to a targeted list, for example. Even so, you won’t get in-depth insights.
Marketers need to use data as an additive, not as the basis for buyer and customer understanding.
Take Advantage of Buyer MVPs to Strengthen GenAI Output
In Forrester’s July 2023 Artificial Intelligence Pulse Survey, almost 50% of organizations said that personalizing customer interactions with GenAI-enhanced experiences will be their most important use case for the technology over the next 12 months.
How’s it going so far?
Isn’t it time to call bullshit on GenAI by arming yourself to know the difference?
A Buyer MVP will help you do that with confidence.
Leave a Reply