ChatGPT – Say What?

The case for toning down your claims to delivering business benefits

A very short history of “features vs benefits”

To start off, I want to make clear that I think it’s excellent advice. Here’s the thinking behind it:

Traditionally, tech companies, and especially those with techie founders, tend to focus on product features in their messaging. And while features are one of the most obvious ways to differentiate from the competition, leading with them in marketing comms brings a whole raft of problems:

Budget: The only people who quickly understand (and most naturally care about!) a feature are other techies: engineers with deep domain expertise. They are not usually the people with the budget who make purchasing decisions (but they may be important influencers)
Relevance: The people who do have the budget are rarely deep subject matter experts – they need help understanding why a feature matters, and why they should care about it. They need it framed as a benefit to avoid a “so what?” reaction
Readiness: At an early stage in the sales cycle – when buyers are gathering information and educating themselves – it’s usually too early to talk differentiation on features. The problem at this stage is winning trust by showing you understand your buyers’ world, and articulating the problem you solve for them (such as delays, or error rates, and their impact on the business). Your marketing job at this stage isn’t to sell – it’s to convince the buyer that they need a solution in the first place. It’s only at a later stage that you’ll show them that your solution is different, and better.

So that’s, roughly, where this best-practice advice comes from: to help techie people frame their product as a business solution rather than a technological one. And to show why your solution matters to more people than just its techie users. That’s why, in messaging frameworks, you often see diagrams like this, that ladder up features to benefits (and align them to buyer personas):

A diagram showing a scale of features to capabilities to benefits to big benefits — This is a very simplified model. For complex products and organisations, you’ll often find many more steps – but it serves to illustrate my point.

The credibility issue

One of the key strengths of this approach has to do with credibility: buyers don’t generally question a vendor’s ability to deliver on a feature, but they may be highly sceptical of lofty claims to a big-ass business benefit (or BABBs, as I lovingly call them). So showing how a feature ladders up to a benefit is a good way to give substance to your claim.

But, I wouldn’t be me if I didn’t have some reservations about the validity of this as universal best practice. As I said above, I don’t think this advice is for everyone.

The case for modesty

I believe that for a significant number of tech vendors and their products (and especially slick digitally-first startups that haven’t been founded by techies), it pays to be modest: first of all, because claims can be tenuous if not backed up by really solid data (eg “improved customer experience” means nothing if you don’t have the NPS score or something similar to prove it); and secondly, the higher you aim with your BABB (e.g. “Competitive Advantage”), the more abstract you become – and risk claiming everything, and therefore nothing at all. So you need to decide which sort of selling personality you want associated with your brand: data-driven or visionary (I’ll write something about this soon).

And I can think of a few concrete cases when your messaging definitely shouldn’t go all the way to C-Level benefits:

When the C-Suite isn’t part of the purchasing decision. I know many clients don’t like to hear this (everyone wants to target the CEO these days!), but be realistic: if you’re not selling mission-critical enterprise systems that can seriously claim strategic relevance, you may have to accept that the CEO will never know that you and your product exist. They may completely delegate the buying process to their IT, user, and procurement departments and ultimately only sign off on the purchase.
When you sales team can’t deliver the story: Features-to-benefits messaging is about getting the altitude right and understanding the complexity and nuances of each department involved. A really great sales person can tell the right story to each customer persona AND connect it all to a big business goal, too. It’s a rare skill and not all sales teams can deliver on it. If they can’t, aiming for something like “competitive advantage” in your messaging is likely to feel like overclaim or fluff to your prospects.
When you’re in danger of neglecting the core buyer. It’s usually not a good idea to approach an engineer with a big business vision (unless they’re the sort of ambitious go-getter every marketer hopes for but rarely encounters); just like you’re likely to bore a CEO stiff with talk about APIs. On a recent project I worked on, the core buyer was a Head of Digitalisation for HR. As I interviewed them about their challenges, I kept probing for things like “retention rates” and “employee satisfaction”, but they kept talking about inefficient processes. The lesson: If I were to try to sell anything to this person, I’d have to frame it in terms of process improvement, not something bigger and more abstract – or I’d probably lose their interest.

Focus on your best prospect and empower them instead

In all of these cases, it’s a much better strategy to optimise your messaging for your core buyer persona (e.g. the department head) and focus on the benefits and KPIs they care about. I’d recommend having your C-level messages in the back pocket and bringing them up in one-on-one sales conversations rather than in your marketing materials.

And: it’s always a good idea to create a piece of content that enables your champions to sell internally – such as

An overview of talking points broken down by buyer group functions
A business case template
An interactive ROI calculator

Or maybe they’ll ask you for a specific format they know their decision-maker will like. Indulge them and have your marketers fix one up. Having an internal champion deliver your message in your stead to the C-Level is usually much more successful than a direct sales pitch.

TL;DR: stay modest, retain credibility, remain relevant

I want to emphasise that modesty is different from lack of confidence: as a tech brand and product, you should definitely know where you play and confidently claim your space in your market. That includes knowing how you can drive business success for your customers.

But in many cases, the real skill is identifying the core decision-maker first (hint: it’s probably not the CEO!), then aligning your claims to their goals and KPIs. When you do that, good things will happen:

You’ll speak to their needs and pain points
You’ll credibly show how your solution makes their life easier
You’ll demonstrate how you can positively impact the KPIs they themselves are measured on
You’ll forge a more natural and credible path to the decisoino-makers

By convincing the key person thoroughly rather than doing a half-arsed job across the board, you ultimately stand a much better chance of selling. This is quiet confidence and it goes a long way.

ShmatGPT

What ChatGPT is great for: email invites

Look, you probably don’t need another hot take on whether marketers should be using AI-generated marketing content, by a marketing strategist, but here it is anyway:

I rarely hear anybody talk about the key assumption behind this tedious debate: the assumption that it’s a marketer’s job to produce a lot of content.

When in fact, it’s the job of marketing to produce good content: ideally genuinely new ideas, presented in a fresh way, that resonate with the people you’re targeting.

From what I’ve seen so far, that’s the opposite of what AI does.

What, actually, is the job of Marketing?

ChatGPT regurgitates things that already exist out there, in ways that already exist out there. Assembling words with the highest likelihood of being used in collocation by the highest number of people. Sometimes, the writing is tolerable. Often, it presents utter shite with a level of confidence I can only aspire to. (George over at Radix has written a fine blog about “AI eating itself” with much more nuance than I can muster).

The output from this exercise is the definition of middling. The literary world’s equivalent of brown. If that’s what you think your marketers’ job is, then go ahead.

In my mind then, you have a completely different problem: that you don’t understand the de-familiarising power of unique words (or images, if you’re using AI for visuals) to help us see well-known things in a new light, or the function they have in engaging audiences, or the blank-everything-else-out effect that a genuinely new idea can have, or the responsibility you have as a brand that targets people to not bore them stiff. And yes, your responsibility to hire and train marketers that understand these things. Who can tell a good headline from bad, who know what a differentiator is, who can brief in and feed back on copy.

If you do understand these things, and still use ChatGPT, then you know that getting the prompts right takes such a level of customisation and tweaking that you might as well use that time to write the damn thing.

A ChatGPT experiment

So far, there’s only one use case I can see where ChatGPT being genuinely helpful, and that’s for functional content that follows a clear template. Invites, sign-up forms, things that have a formula. I’ll give you an example of an instance where ChatGPT has impressed me (and, in red, the bits I loved and the ones where I think it falls flat):

Prompt:

Write a sales email to a prospect in which I invite them to sign up to our webinar on the topic of best practice in Modular ERP. [simply because I had a chat about AI-generated content with an ERP client of mine last week]

What ChatGPT came up with:

Subject: Invitation to Webinar: Best Practices in Modular ERP

Dear [Prospect’s Name],

[it’s great that it signals all the core bits of information that need to be there, such as subject lines and placeholders for names, dates and times. You wouldn’t believe how many emails go out with the essentials missing. Wish it had given me a preview snippet for the email as well, and maybe a second subject line for A/B testing, but it’s pretty good.]

I hope this email finds you well. I wanted to reach out to you because I believe you would find great value in our upcoming webinar on “Best Practices in Modular ERP.”

[“Find great value” is terribly generic. Why/ why for me/what’s the value? Similarly, “finds you well” is tired language, plus double use of “find”. A good editor would spot that.]

In today’s fast-paced business environment, having a flexible and scalable ERP system is crucial for optimizing operations and staying competitive.

[“Today’s fast-paced environment”. I couldn’t believe it went there. The most telling sign of lazy writing. It says you should care but I don’t. It claims urgency without proving it. Absolute no-go. I actually have a voodoo doll in the likeness of that phrase that I stick pins in at night. But: I was impressed that it hit a crucial business benefit of modular ERP. Optimising operations is indeed something a well set up ERP does and I hadn’t briefed that in. I assume that means it’s what everyone says, though. “Staying competitive” is the same as “fast-paced environment”: it’s so generic a benefit – duh – that you should step away from it].

Our webinar will delve into the benefits and strategies of implementing a modular ERP approach, allowing businesses like yours to tailor their systems to meet specific needs and scale efficiently.

[My main issue with this paragraph is that it sounds good at first reading, and then you realise it doesn’t actually say anything. Meet specific needs indeed].

During the webinar, our expert speakers will cover:

The advantages of modular ERP over traditional monolithic systems [Loved that word. I suspect everyone uses it or it wouldn’t have made its way into this]
Key considerations for selecting and implementing modular ERP solutions
Real-world case studies highlighting successful implementations
Strategies for maximizing ROI and long-term sustainability [Great work if you can get it, but I hadn’t briefed that in and I’d want to check that I can actually deliver on case studies and ROI before I send this out]

Date: [Insert Date] Time: [Insert Time] Duration: [Insert Duration] [thanks for the reminders, this is useful!]

This webinar presents a unique opportunity to gain valuable insights and actionable strategies directly from industry experts. Plus, it’s completely free to attend. [Love all these AI-generated case studies and experts I don’t have, that really makes my job a lot easier. Wait, what do you mean, free. IT’S NOT FREE. by which I mean, you really gotta check the content of this stuff]

To secure your spot, simply click the link below to register: [Insert Registration Link]

If you have any questions or would like more information, feel free to reach out to me directly. We’re here to support you on your journey towards optimizing your ERP system.

[Switching between “I” and “We” is not a great look, but I do it on this website as well, so I’ll let it go]

We look forward to having you join us for this informative session. [the rule is “Show don’t tell”. That’s for them to find out, not for you to claim.]

Best regards,

[Your Name] [Your Position] [Your Contact Information]

ChatGPT Really isn’t Marketing’s Biggest problem

Okay, so on the whole this wasn’t bad.

I found it too long and generic, but part of that was down to the minimal briefing. If you’re writing a generic format like an invite, my take is you could save yourself a lot of time by getting ChatGPT to write it for you. But only if you then invest the effort into making it good, and importantly, yours. Saying things that are different and new, and unique to your brand. Also maybe injecting a bit of personality into that voice.

What it alerted me to was this: whenever it came up with bits that sounded industry-specific, but were actually too high-level to say anything much, that was a red flag for me: the way the technology works means that it can only have pulled this from existing content about modular ERP and these things must have been the safest common denominator. They’re the most generic things you could possibly say about the topic, and from my marketer’s POV, therefore useless. They don’t contain any differentiators or value propositions, or product features or audience insight – none of the elements that make marketing good, and specific to you. So I guess you could use it as a tool to establish the absolute minimum baseline of what others in your industry are saying. Then pivot the shit away from that.

My TL;DR point though is this:

It’s Marketing’s job to figure out what to say and to whom

…and ChatGPT can’t solve that for you (not yet anyway).

The real problem is that most companies considering ChatGPT have got the job of the marketer all wrong. They burden in-house marketers with producing all this formulaic crap that nobody reads, and they don’t challenge them to make it good, or give them the time to develop genuine differentiators. They literally think they’re in the business of producing words. If, at a time, where formulaic content production cost has hit quite literally zero, they can’t pivot away from that and start seriously re-considering the job of the marketer, I can only hope the market will punish them.

If, however, it makes them realise that the marketer’s job is to think about all the hard things – what we’ve got to say, why we should be saying it, and why people should care – then I, for one, welcome our LLM-powered artificial overlords with open arms.