Scott Swigart

We're just wrapping up a research project with B2B tech marketing leaders (mostly CMOs) on their use of generative AI for marketing. We haven't done all the analysis yet, but here are some of the clear takeaways. * Companies are in various stages of investigation. Some companies are barely touching it, others are exploring with enthusiasm, and most are somewhere in-between. * No one (that we spoke to) has operationalized generative AI as part of their workflows. Some are getting close, but so far, the direction from marketing leaders is it's more "play with it and share what you learn", or some organized hackathons. * Most marketing leaders are not getting pressure from the CEO or board to adopt this quickly. * Most are using free tools, like ChatGPT. There's low adoption of tools like Jasper or Wryter. * A common use case is to write the "sloppy first copy" of content (blog posts, press releases, etc.) * Generative AI will have a significant impact on external marketing agency spend. I'll have a lot more to share once we complete the full analysis, but no one should feel like their company is way behind if use of these tools is currently limited. If you want a copy of the findings, let me know and I'll email it to you.

In addition to great advice in other answers, there are a few other things to consider: * Most message testing is artificial. People are asked to carefully read and ponder the message before giving feedback. This isn't how people process messages in the real world. There are techniques that can be get at "System 1 thinking" - the automatic and near instantaneous process people will use when encountering your messaging in the wild. This can be as simple as only displaying the message for a few seconds before taking it away. The careful reading is also useful, but later. * Get feedback on critical dimensions. Ask what brands come to mind when they read the message - see if yours is in the list. Test if the messaging is differentiated, or sounds like what everyone is saying. Is it motivating. Is it memorable (check what they remember later in the conversation). Is it credible? Is it relevant? * Pit the messaging against competitors. Show people a mix of your messaging and competitors. Real customers will be looking at messaging from a variety of companies. Does yours win out? All of these things can be rated as on quantitative scales. Hope this helps.

Guided deliberation is a great methodology. You take 4-6 people and divide them into two camps. One camp is pro one solution, and the other camp is advocating for another. You have them debate their positions and work to reach a consensus. It taps into much deeper thinking than what you typically get in interviews and illuminates drivers and barriers.

Foundational research is critical (segmentation / ICP, personas, journeys) and individuals can be found for primary research on just about any topic. With AI and secondary research, this kind of research can now be done at about 1/2 the cost and time of traditional approaches. Once the foundations are in place, a modern research agency should be able to use AI grounded in that primary research to perform all kinds of follow-on research (like message testing), in days not weeks or months.

You can conduct a "first 30 days" study with new users as follows: First, identify potential audiences who can provide feedback on your new feature. These are: * People who are entirely new to your platform. * People who have just started using the new feature. The first group is generally easy to identify since the customer signup date is logged. For the second group, there are a couple of ways to identify them: * If your product logs telemetry on feature usage, you can pinpoint users who have recently started using the new feature. * Alternatively, survey a significant portion of your customer base, asking when, or if, they began using the new feature. Once you've identified your audience, initiate the "first 30 days" study. Offer incentives, such as Amazon gift cards, to encourage participation. Inform participants that you'll be checking in with them weekly over four weeks, using discussion platforms like Slack or Remesh. Each week, present a series of questions or tasks. Utilize both multiple-choice and open-ended questions. Encourage participants to engage in a public channel, allowing them to build upon and refine others' responses. For more in-depth insights, you can directly message participants. After the four weeks, conduct focus groups or one-on-one interviews to gather more detailed feedback. The longitudinal approach of this study yields richer feedback compared to standalone interviews. It also promotes consistent feature usage, ensuring well-informed feedback.

Oh man, the most impactful thing an organization can do is accept bad news. Think of how lucky you are. If you can stop the train, you've only wasted a lot of money on product development. You now have the opportunity to save a ton on launch, ongoing marketing, sales training, sales hours, partner enablement, and more - all of which is doomed to fail if there's no product / market fit. Now it's time for some important questions: 1. What made people think this was a good idea? Something led the product team to think that customers wanted this. Go back to the start and figure out what that kernel was. Is there some merit in the initial idea, just the wrong product development execution? 2. What was the Job to Be Done? The customer likely has a real need, but your MVP just isn't the answer. Understand the need that the MVP was supposed to solve for, because that need is still there? 3. Pivot? It's very common that plans need to pivot. While it's unfortunate that the MVP is almost done, can anything be salvaged from the effort. Disagree and commit. There's no value in sitting on the bad news or sweeping it under the rug. Tactfully and professionally lay out what you've learned, to the people that matter. The goal isn't to shame anyone or put anyone on the defensive but go on record with someone senior (the CEO potentially), about what you've found, and collaborate with them on what to do about it. Even if the MVP moves forward, you've shown you're someone who can be trusted to give what every leader craves - the truth. If the launch moves forward, use the effort to learn as much as you can. Customers are rarely upset that you tried, and they will likely be eager to help you hit the mark in V2. And I don't envy you.