Inside PwC, the idea of a product-led consulting company was something people were cautious of. Jill Foley (https://www.linkedin.com/in/jill-foley-9a6100/) sat in those rooms for 8 years, latterly as a Director in HR consulting.

"We were still in that era of 'No, the magic of this is that we're all so enormously bright. We can tackle any business problem. We don't have to lock ourselves down.'"

She did not buy it.

Her problem with the pyramid was that most client work did not need a large team sitting on site for a year or more. But when that is the model you run, you have to find a way to fit every job to it.

Years later she built her own company, On3 Partners, on the opposite idea: put the product and the intellectual property at the pointy end of your boat. Here is the order she did it in, and what it cost her.

TL;DR

• Start with a problem your clients cannot solve, not with a technology you fancy building. Most consulting companies do the reverse
• Set up the product side as a separate company early. On3 Partners was 2 companies when it had fewer than 5 staff
• Expect to wait about 3 years for a profit and 4 or 5 for scale. Decide up front what you can afford to lose
• Licences now bring in 40% to 50% of revenue, and she can see 3 years ahead. That is what lets her turn work down

Why Jill Foley Refused The Pyramid Model

She has never been a fan of the pyramid, and she started forming that view during her years at PwC from 1990 to 1997.

Most of the problems clients brought did not need a big team on site for 12 months or more. But once you have built a company that depends on those teams, the work has to be made to fit: "If that's the model that you've got, you have to weave that into whatever business problem that you're trying to solve."

And what the consultants build is only ever part of what delivers the result.

"All the knowledge that you build up about the business as you build that solution walks out the door when the consultants leave."

She also had a personal reason.

"That slog of starting to build your revenue each year from zero with no annuity and no security as you look forward, is hard. It's hard on the soul."

Running a big team also kept her away from client work, which is the part she is best at. AI, she thinks, finishes the model off: "with the advent of AI, I think it's just an untenable model."

The UK numbers already show the bottom of the pyramid shrinking.

KPMG's graduate and school leaver intake fell from 1,399 in 2023 to 942. Deloitte's went from 1,700 to 1,400, EY's from 1,800 to 1,600, and PwC's from 1,600 to 1,500, with PwC subsequently lowering it to 1,300. Marco Amitrano, PwC's UK chief, has attributed the reduction primarily to the economic slowdown and its effect on investment and dealmaking, with AI reshaping roles as a secondary factor whose weight he expects to grow.

Those cuts take out more than a cost base. They take out something consulting companies have been selling for decades.

"That's not your value proposition. It's not any longer. I can give you 50 brilliantly sharp young brains to analyse and diagnose this problem."

So if you can no longer sell people, what do you sell?

She built On3 to run consulting income and recurring income side by side, so the company could grow without hiring at the same rate. Here is how, in the order she did it.

Why She Started With A Problem And Not A Technology

Most consulting companies go about this the other way round. They decide they want recurring revenue, and then start asking what they could build.

"It's the wrong place to start. You start from 'what's the unique business problem that nobody has yet been able to solve? Is there a way that we might be able to do that? And what's the technology enablement that that requires?'"

She has watched plenty of companies chase what she calls the nirvana of annuity revenue by starting with that question, and she is clear that the order decides whether it works.

Before you build anything, put it through the question she uses.

What is the problem that no leader would ever answer "I'm not interested in that" to? They might say they cannot deal with it right now. That is a different answer, and it is fine. What disqualifies a problem is a leader waving it away as none of their concern.

Succession was hers. Every leader will tell you they are not confident they have the talent pipeline they need for the future. None of them says they have nothing to worry about. She never had to explain the importance of this challenge.

Then there is the work to leave alone. If 10 other companies already do it, and you have no real reason to think you would be 4 times better at it, stay out. Jill puts that figure forward as a rough threshold rather than a formula, and it still rules out most of the extra service lines a growing company gets tempted by, including the ones you are good at.

The Buyers Came Before The Build

The idea behind On3 did not start life as a product. It started as an attempt to fix a problem at Nortel, where she ran organisation effectiveness through a hypergrowth, heavily acquisitive phase she describes as "more like an extreme sport than a job", and where the usual ways of managing talent were not working.

So Jill worked backwards from the future instead: where does this business need to be in 3 years, what will it need to be good at, and what talent does that mean we need in the pipeline now. She calls that approach Future Back, and it is still the core of what the company sells.

By the time she set up the business, 3 or 4 organisations had already told her the approach looked different and might solve a problem they all shared. She had buyers before she had a build.

“We didn't go out looking for something to technology-enable."

Only then did spending money on technology make sense, and even then it was for one specific job, which was to run the analytics faster.

She had also spent her early career implementing technology, and had been told many times that technology projects fail more often than almost anything else. So she ring-fenced the money, named the figure she was willing to write off, and settled in to wait. Patience and tenacity are what a build like this asks of you.

Splitting The Business In Two

Jill split On3 into 2 separate companies while the business still had fewer than 5 staff.

One of them held the technology, which walled off both the spending and the financial exposure, and she brought in a technology specialist as her partner to run it, because doing it alone would have taken her well outside what she knew.

The consulting company stayed separate for a different reason. The 2 businesses need different governance, different skills and completely different sales and marketing. Sixteen years on, Jill still thinks keeping them well apart is what made it work, and she thinks bigger companies could learn from it.

No holding structure for a future sale, no tax arrangement. Two companies, so that one could not eat the other, set up at a point when the whole business could have sat round a single table.

Why Her Method Leads And Her Technology Follows

Plenty of consulting companies now sell something alongside the day rate. A diagnostic, a benchmark, a platform, a licensed method. So where is the line between a consultancy that has a product and a product company that happens to do consulting? It comes down to which one pulls the other along.

"So what leads?” said Jill. “Is it the technology that creates the consulting revenue or is it the methodology that then opens up the possibility of building a licence?"

Start by asking what the product is. If it is a piece of technology, you now have to build it and keep it running, and that is a different business from consulting. If the product is really a proposition, where you are the company people call about acquisitions or due diligence, you have a different set of problems.

In her own company the order never changes. The method comes first and creates the need for the technology. She sells consulting and a distinctive proposition, and once a client has seen it work, they take a licence. She never leads with the software. That order decides what each side of the business is worth, and it is why the 2 companies stayed distinct instead of blurring into one.

She had a commercial reason for taking the unfashionable side of this argument, not just a philosophical one. When nobody has heard of your brand, you have to stand for something, and the easiest way to do that is to be visibly better than anyone else at one narrow thing.

What The Model Paid Her And How Long It Took

Technology licences now bring in roughly 40% to 50% of On3's revenue, and at any point she can see 3 years of income ahead of her.

"We've got a secure revenue stream 3 years out, which helps me sleep at night. But it also gives us choices as a business. We can be quite choiceful about the consulting work that we take on."

A company that can see so far ahead gets to pick its consulting work, while one rebuilding its revenue every January takes what it can get.

The technology business made no profit until about year 3. It did not reach any real scale until year 4 or 5. Most of the money it has made across 16 years arrived in the later years rather than being spread evenly. That is a long time to keep paying a partner, running a second company and holding money aside for losses, and it is the part that usually gets left out when people talk about packaging up their IP.

What She Refuses To Sell And What She Outsources

None of this works in a company that refuses to narrow down, and the narrowing comes first rather than as a tidy-up at the end.

"My own approach to this has always been to stay narrow, stay deep. And what I've learnt over time is it never, ever boxes you in, in the way that you might fear"

Every consulting company under pressure to grow has the same fear: say you only do one thing and you will be shut out of everything else. Sixteen years in, the work kept coming. Being known for one thing made people buy more easily, because they knew what she stood for and they knew when to pick up the phone. After one conversation, they could tell you exactly what would make them call her.

She runs her own company the same way. On3 has never done its own marketing. It outsourced software development for years and only brought it in-house when the numbers said to. The rest of the work around the edges goes to virtual assistants and partners. The companies Jill sees doing best right now are the ones that have worked out what they are good at and what they are not, and then gone and found world-class partners for the rest instead of building their own mediocre version.

Her filter for sorting the work has 3 tiers.

That middle band she never fully pins down. She frames it as a question about where the work is best done, and mentions AI as part of the answer without finishing the thought. The first 2 tiers are the sharp ones.

Asked how a company heading for 200 or 300 people should apply all this, Jill queried the goal instead of answering it. Why do you need a company that size, and what is it for? If what you want is wealth, revenue or profit, you might well need 200 people working on the job without employing any of them, pulled together from several different places. Your role then is to aggregate, set direction and invest.

A headcount target is common enough to pass for strategy. Ask what the headcount is for.

Does AI Make Any Of This Faster?

The obvious hope is that AI compresses all of this. Products get built quicker, IP gets packaged faster, and the 3 years to profitability becomes something shorter.

Nobody has this figured out yet, and leaders are being asked to work it out while already stretched by how much they are running and how fast everything around them is changing. Thirty years in, Jill puts herself in the same position: "I'm running to keep up along with everybody else, in terms of AI." Then she names the mistake she keeps seeing.

"The obvious misread is that AI is a technology solution and it's an efficiency play"

These models are built on what we already know, not on what we can predict about the future, which limits how much they can help you with a proposition whose whole value is that nobody has cracked the problem yet. The better question is how to use AI to strip out the low-value work so that people can do more of what only people can do. The efficiency then arrives as a side effect rather than as the point.

So if you are building towards recurring revenue, the answer is dull. AI might shorten the build. It will not help with the hard part, which is finding a problem worth solving and proving you can solve it.

Her Question For Anyone Who Says They Have Let Go

Everyone in this industry is talking about the end of the pyramid. Talking about letting it go and letting it go are 2 different things.

"You may already have let go of it intellectually, but have you let go of it emotionally? Have you done the mindset work which says, 'okay, as a partner in this firm, what is the new contribution that I should be making? What does that mean for me and the assumptions about my value and my worth?'"

Two more questions follow, and you cannot hand either of them to anybody else.

1. What is it your job to lead now?
2. What is it your job to stop leading?

Most people answer the first one and stop there.

The small and mid-sized companies have mostly made this move already, which is why they are doing what they are doing. In the big ones, people know the model they are sitting on is the wrong one and have not let go of it yet. Everything else in this article is a structural decision. This one is not, and it decides whether you ever get round to the others.

The Bigger Picture

You can spend a consulting career chasing revenue wherever it turns up. Jill has done it: "I've worked in a system where we chase the money and it's a miserable place to be."

She went the other way. Work out what you are better at than other people, keep getting better at it, give it away generously, and let the money follow. She calls herself willing to be the experiment. She also says it is hard to do inside somebody else's company.

If this only works when you own the place, then every partner who believes it and cannot act on it is telling you something about their company rather than about the idea.

Frequently Asked Questions

How do you build annuity revenue in a consulting company?

In this order. Start with a client problem nobody has solved, not with a technology you could build. Get real client interest and proof that it works before you spend money. Put the technology in a separate company, because the 2 sides need different governance, skills and sales approach. Then let the method create the need for the technology, so you sell consulting first and clients take a licence afterwards. Firms that start by asking what they could build usually fail.

How long does it take for a consulting product or technology to become profitable?

Longer than most business cases assume. In one 16-year example the technology business made no profit until about year 3 and reached no real scale until year 4 or 5, with most of the money arriving in the later years. Licences eventually reached roughly 40% to 50% of company revenue. Ring-fence the money and decide up front what you can afford to write off, because technology projects fail more often than most.

Should you separate your product business from your consulting business?

On3 Partners split into 2 companies while it had fewer than 5 staff, for 2 reasons. The split walled off the money going into technology and the risk of losing it. And the 2 businesses need different governance, different skills and completely different sales and marketing. Sixteen years on, keeping them well apart is what made it work, and the same logic applies to bigger companies trying the same combination.

At what point does a consulting company become a product company?

It depends which side pulls the other along. If the technology brings in the consulting work, you are running a technology business and you have to build and maintain software. If your method is what makes a licence possible, you are a consultancy that has a product. At On3 the method leads: the company sells consulting and a distinctive proposition, and clients take a licence once they have seen it work.

How do you decide which problem to build a product around?

Look for a problem every leader has. Which problem would no leader ever wave away as none of their concern, even if they say they cannot deal with it right now? Succession is one example, because every leader admits they are unsure about their future talent pipeline and none says they have nothing to worry about. Then apply the reverse test: if you have no real reason to think you would be far better than the companies already doing that work, leave it alone.

Does narrowing your proposition limit your growth?

One company that narrowed on purpose found the opposite. Being known for one thing made clients buy more easily, because clients knew what it stood for and when to call. The fear of being shut out of other work never came true. The same discipline applied internally: it never did its own marketing, and outsourced software development until the numbers said to bring it in-house.

Can AI shorten the path to productised revenue?

Partly, and not where it counts. The mistake is treating AI as a technology fix and a way to cut costs. These models are built on what we already know, not on what we can predict, which limits how much they help when your whole proposition rests on solving something nobody has solved. AI might shorten the build. It will not shorten the work of finding a problem worth solving and proving you can solve it.

Is the consulting pyramid model still viable?

The UK numbers show the bottom of the pyramid shrinking. KPMG's graduate and school leaver intake fell from 1,399 in 2023 to 942, Deloitte's from 1,700 to 1,400, EY's from 1,800 to 1,600 and PwC's from 1,600 to 1,500, with PwC later cutting to 1,300. PwC's UK chief puts this down mainly to the economic slowdown, with AI a growing second factor. The commercial upshot is that you can no longer sell a large analytical team as the offer in itself.

Do you need 200 people to build a substantial consulting business?

Ask what the 200 people are for before you commit to hiring them. If what you want is wealth, revenue or profit, you might need 200 people working on the job without employing any of them, pulled together from several places. That makes your role aggregating, setting direction and investing, rather than managing headcount. A headcount target is common enough to pass for strategy. Ask what the headcount is for.

What does forward revenue give a consulting company?

The ability to say no. A company that can see roughly 3 years of income ahead gets to pick its engagements and turn down work that does not fit. One rebuilding revenue from zero every year takes what it can get, which shapes the quality of its work and, over time, what it becomes known for.

About the guest:

Jill Foley, founder and chair of On3 Partners and co-founder of The OD Collective, has spent more than 30 years in organisation design and organisation development, including 8 years at PwC where she finished as a Director in the HR consulting practice and 3 years as Head of Organization Effectiveness at Nortel during a period of hypergrowth and heavy acquisition. She founded On3 Partners in 2010 as her second consulting business, having previously co-founded and built 7days, and she is the creator of the Future Back approach to talent management and organisation design.

About On3 Partners:

Founded 2010, UK-based, specialising in organisation design, talent management and the technology that supports both through its Future Back methodology. Roughly 40% to 50% of company revenue comes from technology licences, with a secure revenue line 3 years forward. Clients have included Mars, Novartis, Cathay Pacific, Sky, Bupa, Whitbread, Google, PayPal, Inchcape and BBC Worldwide. Jill Foley's related venture, The OD Collective, co-founded with Carla Henry, teaches the 8-step Circle Line organisation design methodology to senior leaders.

This article is based on an episode of the LEADERS IN CONSULTING Podcast, hosted by Sammy Gebele, Founder of SAWOO.

If you want to hear the full conversation behind this analysis with Connor, you can find the episode in the podcast section.

If you’d like to stay in the loop on future deep dives and community updates, you can subscribe to the newsletter here.

For more perspectives on how leaders are navigating challenges like this, follow LEADERS IN CONSULTING on LinkedIn.