Does AI for B2B Marketing live up to the hype? That's the question I'm exploring in this three part series.
In part 1, which you can find part 1 here, I asked AI to suggest Ideal Customer Profiles for our product, CMO Copilot. The upshot is that the ICPs AI suggested mapped pretty closely to the ICPs we developed on our own, using good, old-fashioned surveys, interviews and analysis.
There were some differences, as you'd expect, but overall, the AI-generated suggestions would have sent us in the same direction we actually took after doing our initial market research.
The only real surprise was the rationale it provided for focusing on eCommerce professionals, which we didn't think would be a good, initial target for our product. (I'll follow up on that in future articles.)
Now, in part 2, I'm using those basic suggestions to generate complete ICPs.
I should point out that I'm a big believer in planning, so I use ICPs as the foundation for everything I do in sales, marketing, design and even product development.
The reason for doing this in marketing and sales is that the modern B2B buying journey is so non-linear, the only way to do content marketing (IMHO) is to meet customers where they are, when they are, by creating the kind of content they want, and putting it where they can find it, when they're actually looking for it.
(In design and product development, I just think it's easier to manage workflows that are organized around ICPs, and I believe it helps that they're all aligned so you can see the sales and marketing effort that's associated with UIUX specifics.)
While this is definitely a more front-loaded process, I think it saves 2x to 10x in time and effort over the long run, and doing it this way is much easier to manage, revise and refine, too.
Test Results, Pt 2:
Over two prompts, AI suggested a total of eight ICPs. In my experience, most B2B businesses do well with one to three ICPs, sometimes three to five ICPs, and occasionally five to seven ICPs.
I have worked on a couple of professional services projects that had 15 ICPs, but those were unusual situations (a professional services company and a public works project) where there really were meaningful differences between that many target segments (which we derived from a combination of specific use cases and jobs-to-be-done, if you're curious).
For testing purposes, I'm going to focus on the three ICPs which have broad use cases for our product.
1.
This is what AI suggested:
Profile: Freelancers & Consultants
Age: 30-45
Occupation: Digital Marketing Consultant
Key Adjectives: Independent, innovative, flexible, determined, tech-savvy
Rationale: They are constantly searching for tools to enhance their offerings to clients. The "CMO Copilot" will be useful since it automatically generates personalized content, and can adopt the persona of their different customers, a feature that would streamline their task of juggling many clients.
Here is the ICPs we developed from that suggestion:
2.
This is what AI suggested:
Profile: E-commerce Leaders
Age: 28-40
Occupation: E-commerce Manager
Key Adjectives: Competitive, strategic, tech-savvy, results-oriented
Rationale: More than others, e-commerce managers often have a dire need to drive engagement and maintain buyer interaction. The AI for marketers and behavioral intelligence feature in the product would help them achieve this while maximizing returns.
Here is the ICPs we developed from that suggestion:
3.
This is what AI suggested:
Customer Profile: Digital Marketing Specialist
Age: 28
Occupation: Digital Marketing Specialist at a startup
Adjectives: Tech-savvy, ambitious, competitive
Rationale: They constantly need to provide fresh and engaging content in a crowded market. The feature of automatically generating personalized content in CMO Copilot can lessen their workload and ensure they maintain audience attention.
Here is the ICPs we developed from that suggestion:
Useful?
I certainly think so, but, as I said earlier, I'm in the "measure twice, cut once" school of marketing, so taking a little extra time to think and plan is always preferable.
Plus, these ICPs reflect everything our team believes separate good B2B ICPs from bad B2B ICPs — professional priorities and pain points, company KPIs, various behavioral tendencies, as well as the key strategic and tactical recommendations that arise from those like messaging voice and tone, which kinds of subject lines and content to use, what "first contact" might look like, etc.
Good ICPs ---> good strategies ---> good outcomes.
In part three, I'll use these ICPs and the included B2B recommendations to generate both marketing management materials (creative briefs, marketing calendars, industry research, etc.) and various types of personalized content (emails, sequences, headlines, etc.).
Comments