How to Implement MarketSauce: Masterclass Follow-Up (August 2024)

Full video transcript

Good morning everyone, and welcome into the MarketSauce Method Masterclass follow-up call number one here on August 8th, 2024. I am Darby Rollins, joined today by an amazing group of people including Shelly, Roger, Marty, Marque, Tyrese — great to see you all here — and Yolanda. Happy Thursday everyone, good to see everybody here on today’s call.

The agenda for today’s call is to first and foremost start off with any wins that anyone has coming off of the masterclass and workshop last week, insights or anything specific that you took away and have been putting into action. We also want to make sure that we’re addressing on this call any particular roadblocks or areas that you’d like us to maybe mastermind or hot-seat during the call, if you’re stuck anywhere particularly putting things into action. Then we’ll also be discussing a few things related to the certification and just general feedback as we continue to move forward.

So we’ll open the floor first and foremost: are there any wins coming off of last week’s masterclass that anyone would like to share? Tyrese went first: “I was super impressed by the fact that I went through the whole thing, and I thought because we’d already done the MarketSauce 9000 there wasn’t more to discover, but with this it really dove in deeper. I managed to get 78 pages out of the entire prompt.” He ran through all the prompts and copied and pasted the output into a Google Document to see where all the answers went, then changed the inputs to see how the output would differ for a course versus another service — “it’s giving lots of new ideas.”

Marque shared a win with a new client and lead: “I used all the input of the last MarketSauce document, so I started with questioning my customer, put the questions to them, got the answers, put those answers into the persona development — then there became a persona, it was a team persona, and they asked us to facilitate setting up an innovation hub in their organization.” Once the persona was clear based on their input, Marque mapped the service offerings to their question and became aware that, based on the prompting and structure, it became “very crystal clear” what the key benefits were for that persona specifically. The intake call — a Zoom meeting — became “the best call ever” because Marque had the best information ever, and by the end the prospect was asking for an offer. “My lesson here is first ask them questions, get their answers, because then you already have some real data from your customers — add that to ChatGPT, in my case, and then develop the persona and map your service offerings.”

Another member described doing market research on older men in the later stages of their careers: “We had an intentional dinner the other night — we had eight guys come together, most of us didn’t know each other — and everything they said was in the reports that I created, like what they want out of life at this stage and what their pain points were. It was just amazing.” The reports were generated by starting from personal reflection, talking to a few friends, and then validating with the larger group — “it’s just really powerful to see how spot-on it was.” Darby noted this builds on the prior week’s conversation about AI focus groups versus real-world human focus groups and matching up what’s true.

Marty shared a win that didn’t close a client but still helped: through the conversation, someone had suggested focusing on one market rather than coaching across many verticals. Marty picked healthcare, started building content, then realized a deeper understanding was needed and called a surgeon friend, who got excited going back and forth on the document. The surgeon’s real-world grounding surfaced a concrete idea — post-surgery care instruction sheets that nobody reads but are required by regulation — as a use case AI could generate and email automatically, saving time and creating a record it was sent. Marty also considered shifting the offer toward a role-based angle (e.g., product managers across industries) but hadn’t started that yet. “All that came from having to focus on one thing, which helped a ton — the information I got out was so much better.”

Another member described a meeting with a friend who is a lawyer doing high-end tax planning in Canada using life insurance strategies. Basic MarketSauce input produced good information; the client pointed out what was right and wrong, then shared a paper he had written that could be uploaded to refine the avatar questions. Despite the “boring” subject matter of life insurance and actuarial science, the value proposition is significant — potentially reducing a client’s tax rate from 60% to 15% in Canada.

Roger shared a smaller win: getting all the way through the MarketSauce GPT and prompts for the first time, which surfaced some functional questions for later in the call.

Kelly described spending more time than usual re-reading the document and combining different prompt components to see how results change, working to use the material “to its fullest extent” rather than just running it once. Darby responded that everyone in the group is applying MarketSauce a little differently, and that’s part of the value of generative AI for entrepreneurs — recognizing that one person’s workflow won’t work for everybody, and that a “cookie cutter” approach is less effective than finding what fits your own business.

Darby also shared a win: working with a client whose MarketSauce application is focused on client discovery for investor-facing materials. The goal was avoiding overwhelming the client with a 50-70 page document by extracting only the 10-15 pages of insights that provide the most value, referencing specific citations back to the underlying documents (e.g., “in slide 24, the brand references this”). The broader takeaway: discovery calls consume real time (4 hours a week, 40 hours a month) that can be compressed with blueprints, adding value beyond the discovery conversation itself.

Shelly added that going “narrow and deep” on a target market (baby boomer women) created a convergence of energy and ideas — a “Boomers in Business” podcast, a completed speaker page, and applications for podcast guesting. Ted (referred to as Ted in the discussion) shared results from testing ad phrases aimed at a men’s version of the same audience: six candidate phrases were tested via Facebook ads, with “men who fear isolation” getting 10x more clicks than the others, using about $20/day and $60 total spend for roughly 18 leads, following a format learned from Chris Roso around the four stages of retirement.

Before moving to Q&A, Darby gave a brief update: a certification process is being built and should be ready within a few weeks, involving roughly 30-50 questions and about 30 minutes to complete, covering the MarketSauce method and the workflows from the masterclass. The team is also planning a separate, lower-priced membership (apart from the main Gen AI University community) for ongoing follow-up calls — roughly two calls a month — plus community interaction between calls and continued resource updates.

Roger opened the Q&A: earlier in the week he had a discovery call with a potential client where he didn’t know much about the business going in, which raised the idea of a prompt that could help generate discovery questions and verify understanding of a business before and after a call. He then asked a functional question: in the MarketSauce Game Changer GPT, when a user prompt isn’t filled in with real information, does the GPT manufacture information based on what it already knows? Darby confirmed that yes, without real input the model will often fill in context it already has, and recommended uploading real client information (an interview, a discovery call, existing documents) before running the prompts, or handling research separately to avoid lengthening the main thread and degrading response quality.

Tyrese shared a related experience: after creating a MarketSauce 9000 document, he used it while job-hunting by asking ChatGPT what questions the document was answering, then used that framework to research prospective employers before interviews. In one interview, the interviewer asked how he knew so much about her business — information not even on the company website — and he admitted using AI, including a 26-page document with a SWOT analysis and gap analysis prepared beforehand. He advanced to a later phase of the interview process. He cautioned the group not to give away that level of work for free but also not to withhold value out of scarcity thinking. He also shared that the content style shared by Ted is well suited to LinkedIn polls — running a poll early in the week, then following up with a newsletter or article responding to the poll results — and that thoroughly reading a prospect’s LinkedIn profile and website before a sales call is a differentiator he relies on personally.

Marque asked about using LinkedIn content performance data (engagement, format, publishing time) as user input to see whether it would make new content output more consistent in quality. Darby agreed this is worth building into the workflow as a future module — analyzing published content for trends and engagement to identify pain points or themes worth doubling down on, focusing on the roughly top 20% of best-performing content rather than only generating net-new material, particularly for someone posting three times a week on LinkedIn for years.

Roger asked a follow-up about the software product itself: previously Darby had generated a “brief lite” from just a few paragraphs of input, which was already fairly comprehensive — how does that compare to running the full MarketSauce GPT with all its user-input questions, and where does the in-development software fit between a light brief and a comprehensive brief? Darby explained that the first workable version of the software will let users submit orders (using credits) to generate that lighter 10-15-20 page version through a self-service flow — company info, a blurb, possibly an attached file — with the ability to place multiple orders organized inside an account. A second stage, still being figured out, would let users upgrade or evolve an existing blueprint over time based on new context, eventually supporting a chat-style interface for talking to the “avatar” and continuing to build on existing content. The initial build is on the OpenAI API, but Darby was clear the intent is not to lock users into one platform — content and blueprints should be portable so they can be applied inside ChatGPT or Claude (Darby mentioned still using Claude 3.5 Sonnet regularly), rather than trying to build an all-in-one CRM-style platform.

A member based in the Netherlands raised a data protection question under EU law: in some cases it’s not permitted to use free versions of tools like ChatGPT or Claude with customer data, since providers may use that data for training. She asked how the MarketSauce platform would protect customer data, noting her own business currently uses ChatGPT Team and Microsoft 365 Business Standard with Copilot for this reason. Darby responded that customer data would not be used to train models, that storage and databases would be kept independent per account, and that enterprise-level integration details would need to be confirmed by a colleague (Chris) who handles that side, but that the intent is enterprise-grade data protection throughout.

As the call wrapped up near the top of the hour, Darby confirmed a follow-up call would be scheduled for roughly two weeks out, with email invites and calendar entries to follow, along with updates shared inside the Circle community and via newsletter in the meantime.