Q&A: AI, Founders in New Biz, Failed Service Offerings, Narrowing an ICP

These are collected from a LinkedIn post that invited you to submit questions. If you have a question, please send it along and we'll do our best to answer it.

Thanks so much for continuing to host these occasional Q/As. I have a question regarding AI tools: is there a benchmark metric for ongoing AI-related expenditures that agencies can track and include in their financial dashboards? —Mohamed Abdelmenem

None of the AI tools are making money, and that’s because they are in the adoption phase and not the profit stage. That means that they are providing far more value than the subscription price, so I wouldn’t worry about that at the moment. Besides, just pretend to cancel a tool and see what deeply discounted price they come back with, immediately.

I think we should still be in the tool exploration phase, which means not settling in on one and instead spending money on multiple. I was a big fan of Perplexity and now the wonder is gone. Claude is much better for checking my writing for non-sequiturs and typos. ChatGPT and Claude, when fed identical prompts, keep changing on the leader board. In a year or so I suspect that we’ll see massive changes in the market and Gemini will likely have won. When that happens, there will be company-wide pricing and appropriate fences around proprietary information, while still allowing the “memory” of the entire organization to yield better results. I’d bet on the price being <$150/person in bulk.

In summary, don’t worry about the money. Experiment with all of them concurrently. The cost of AI is not material at the moment.

How does “Leaders very visible and driving new business” play out in the context of M&A? Are potential buyers skittish about leader-led lead gen? —Adrian McIntyre, Ph.D.

The truth is a bit nuanced on this one. Buyers want a lead generation system that can be measured and can scale (that ain’t referrals), but they actually want sellers to be involved primarily in new business, and the leaders who give the buyers confidence that they are good at it will already been doing it. The idea of the earnout wasn’t originally invented to manage buyer risk, but rather to identify financial obligations that the buyer’s accounting and investor relations people needed to account for and disclose to the public. It later morphed into a way to derisk a purchase, and in that light, the buyer will always consider the biggest risk to be ongoing new business (they believe they can fix everything else). Even the risk of client concentration can be mitigated with effective new business.

There’s another really interesting facet to this: the more “famous” the seller is, the more such a transaction makes sense, and the more the earnout will be derisked by the buyer (because new business activity will be additive).

There’s an element of this, though, that hardly anyone ever talks about: that seller notoriety before and after the transaction will always help that same seller move on to his or her next chapter when the earnout is over. Have you always wanted to write a book? Account for that in the terms of the earnout, and the fame from that book will help the buyer, now, and the seller, now and later.

If you’ve launched a new service, got lots of interest but few bites, what would be your process to diagnose why it didn’t land (i.e., is it the price, messaging, name, expected outcome, etc.)? —KH (name withheld by request).

I like this question because it’s the springboard to get philosophical for a minute:

• Most of your clients should buy most of your services most of the time (see article on service offering design).
• That means minimal adoption is not the goal, and if very few of your clients are buying that service, you really do want to evaluate things sternly, because the goal is much higher adoption. Your goal is to either figure out what’s wrong or to quickly reverse your mistake in overestimating marketplace interest.

Here are some other points to consider:

• Most new service offering initiatives should come from current clients and not new prospective clients. The best outgrowth of a new service comes from a client who already trusts you and wants you to “do this thing too” for them. Be really careful hiring people based on promises from new clients.
• Never start a new service offering unless you fully intend to do your best to sell all those services together in a new business setting, including the new one. Think Prix Fixe menu rather than a cafeteria where clients drive the purchase.

Now let me suggest where you might tighten or loosen these standards. I’ll note first that one of the easiest ways to lose money in a firm is to have a team that’s dedicated to a specific service offering just sitting around with nothing to do. You could lay them off, of course, but there’s always that elusive carrot of wanting a team ready when somebody does bite, and that instinct usually rules the day.

Where you might take more chances on a service offering is if one of these are true:

• The team that will staff this are largely interchangeable with other service offerings. Think UX and front-end engineers or video shooting and editing, but not skills that are rarely found in the same person.
• You can use contractors and only pay them when you have the work.

Back to your question: yes, it could be any of those things you mentioned, but I’d urge you to make sure you should have started offering it in the first place.

What is your thought on narrowing down a professional services firm’s ideal client profile (ICP)? —Peter Caputa

I’d want that firm to follow the standard recommendations set out in The Business of Expertise, for sure. That includes 10-200 competitors, crowding the lower end and at least 2,000 prospects, but more toward 10,000 (see detail here). And the others covered in this article on the five pre-tests and the four post-tests.

But I’d like to add one important component that has become more important than ever in a very fragmented world: addressability. You can pass all those nine tests and still not have an “addressable” market. The easiest way to check yourself on that is to imagine a LinkedIn campaign. Can you isolate very specific targets that you know will meet that criteria? Here’s an example of one you could: a specific C-level title of a firm in an NAICS code for firms with revenue of $100m-$1b. Here’s an example of one you could not: PE-backed firms who want to buy UX research and coding in the same agency, or companies that desperately need internal alignment before they waste money on external campaigns. And so on. And by the way, this does not always require a vertical specialization. There are ways to target horizontal ones.

When a client relationship goes south due to a mistake, are we still able to get back on the authority horse? Do we lose all credibility at the first mistake? I guess I am asking if it is worth it to try to mend a relationship based on aligning values which might be greater than a mistake? —Luis Ramos

Just honestly, there’s no room for a “values” discussion at this point. Your values are your values, and whatever they are has been very apparent to the client. If you go to establish/declare values at this point, it’ll backfire. I wouldn’t even talk about them.

But can you rescue a relationship after one big mistake? Absolutely, and it’s always worth a try. Just be brutally honest (vs. defensive), put yourself in their hands, and live with the results.

Here’s a very important point to remember about this: the presenting issue is often not the real issue. So the client might read you the riot act over a specific issue, but often they do that because that issue is just a lot easier to talk about then the real reason, which might mean disrespecting them, making them look silly in front of their boss, embarrassing them with a missed deadline, etc. So always look for the real reason behind the stated reason, and absolutely fix it. If you’d done a great job leading the client with honesty and a true concern for what’s in their best interest, you’ll absolutely get a great chance to salvage the relationship. Probably not two, though. Once you commit that second big one, they’ll fire you in their head…and let you know about it later after they’ve lined up a new firm.

I left one question out of this because it was so interesting that I wanted to devote an entire article to it: “if you would start a professional service firm from scratch, how would you do it and why?” The answer to that comes next.