Can you buy a collection of small, profitable internet businesses and end up with one large, durable company?
I spent five years finding out from the inside. I was CEO of Orbit, a company that built and ran several Shopify apps, and Orbit was one of the companies owned by Tiny. My job was to grow the apps that deserved it, decide where more investment was justified, and close the ones that no longer earned their keep. Along the way I evaluated about a dozen potential acquisition targets.
The roll-up model makes sense to me. Venture investors back unproven startups and hope a few pay for the rest. An acquisition-led holding company buys businesses that already have customers, revenue, and profit. It owns them, allocates capital across them, and uses their cash to buy more, reinvest, or pay shareholders.
Plenty of good software businesses will never reach venture scale, and this model gives them a home. Buying an established company also removes a whole category of uncertainty. Customers have shown they will pay and the company has proven it can make money serving them.
What stays uncertain is how long those profits last and what it costs to keep them. Competitors erode an early lead. Technology changes what customers expect. Growth may depend on finding a repeatable way to acquire customers at a cost the business can afford.
Those risks shape what the holding company should buy, what it should pay, and how much cash it can expect back. But they do not stop moving once the deal closes.
That is where I think the model faces its most revealing test: can the holding company and the operator running the subsidiary change the plan when the business no longer fits the assumptions that justified buying it?
Some subsidiaries stay healthy with steady investment. Others need a serious renewal, have a real chance to expand, or are better suited to harvesting their remaining profits. For the holding company, durability means keeping the portfolio’s earning power at worthwhile returns on capital while all of that changes underneath it.
Different strategies inside one subsidiary
Orbit’s apps did not share a competitive position, an investment need, or a growth ceiling. Each one needed its own decision, and for five years those decisions were my responsibility.
Some of the apps we inherited carried technical debt. All of them needed ongoing work to stay compatible with Shopify’s changing APIs, platform features, and deprecations. That work kept the apps running. It did not give a merchant a stronger reason to pick us over a competitor.[1]
For one legacy app, I did not believe a major rebuild would earn an adequate return. We harvested its remaining profits while continuing to support customers, and directed investment elsewhere. We also sunset products whose prospects no longer justified the cost and attention they consumed.
Other apps deserved more development and more sales and marketing effort. For those, I had to weigh the likely return against the resources required and the cost of waiting.
Timing belongs in that calculation. While a team spends two quarters on a planned improvement, competitors keep shipping and customers keep comparing. Stretching a project out lowers the spend this quarter, but it can also miss the opportunity the project was meant to capture.
We found one new opportunity by watching customers. Merchants were bending one of our apps to solve a problem it was never designed for. We interviewed them, studied their workarounds, and researched the alternatives. The need was real enough to justify a dedicated product, so we built it.
The overlap mattered. We could reuse product concepts and technical experience we already had. An unrelated opportunity would have been a much larger bet and would have pulled attention from the other apps. We chose a problem worth solving that the team could realistically deliver.
So inside one small subsidiary I was growing, harvesting, closing, and building new products at the same time. The useful distinction was never good product or bad product. It was the return each course of action could offer, given where the product stood and what it would consume.
A subsidiary can outlast its original product lineup. The customer knowledge and technical capability it has built give the team a head start in developing its next source of revenue.
How painful would it be to switch?
Two profitable software companies can report the same revenue and margins and carry completely different retention risk.
Take two hypothetical products. One adds an optional feature to an online store. If a competing app does it better, the merchant swaps it out with little disruption. Unplug one, plug in the other. The second product runs a veterinary practice’s daily operations. Replacing it means migrating years of records, retraining non-technical staff, rebuilding integrations, and changing how the whole clinic works.
For the merchant, switching takes an afternoon. For the vet practice, it takes a planned transition across the business, and a competitor has to offer enough improvement to be worth that pain.
Constellation’s acquisition criteria are built around this. They favour proprietary B2B software with low customer attrition and recurring revenue, mostly serving a specific vertical or geographic market. There is no minimum revenue. A business with a handful of employees qualifies if the economics fit.[2]
Veterinary software is not a made-up example in that world. Constellation’s operating group Jonas acquired VetlinkPRO in 2022, a practice management product that covers everything from appointment booking to laboratory management.[3]
Mark Leonard addressed this directly in his letter covering 2013. Constellation had bought a number of higher-churn businesses, and he separated churn caused by turnover in the customers’ industry from churn caused by low switching costs. The first kind can be fine. The second, he warned, tends to produce businesses “requiring tremendous effort to stay in much the same place.”[4]
I would not draw the line at Shopify app versus vertical software, though. A Shopify app can become deeply embedded in how a merchant runs the store. A vertical product can bleed customers when bad support or an aging product finally makes switching worth the trouble. The industry matters because it shapes how customers use the software, what alternatives they have, and how disruptive a change would be.
For an acquisition, the practical questions are how painful switching would be and how much ongoing investment it takes to keep customers choosing you. Two businesses with the same profit today can leave very different amounts for distribution after that work is paid for.
Profitable does not mean ready to scale
Retention is only part of it. A profitable company may know how to serve its existing customers without knowing how to find many more of them at a price that works.
Picture a business that has grown on referrals and organic discovery. The founder has spent almost nothing on marketing. That is real evidence of demand, but it does not establish whether paid ads, a sales team, or a content program would bring in customers at an acceptable cost.
The next stage of growth might be attractive. It might also take months of testing to learn which channels work, what a customer costs, and whether those customers stay long enough to repay that cost.
If a holding company buys the business for its existing cash flows, it can underwrite it on that basis. If the price or the growth plan assumes that more marketing produces more profit, the untested economics of that expansion belong in the investment case.
“Profitable today” and “ready to grow with more capital” are two separate conclusions.
A durable portfolio does not need permanent subsidiaries
A holding company can buy businesses with finite lives and harvest them. At the right price, after allowing for customer obligations and eventual wind-down costs, that is a perfectly viable model.
The proceeds can go to shareholders. But if the ambition is to sustain the holding company’s earning power, something has to replace the businesses that decline: more acquisitions, growth elsewhere in the portfolio, or new products.
Replacing those earnings takes capital, often a lot of it. A flat total profit figure does not prove an attractive return. The return has to be judged against the capital spent to hold that profit flat. The prices paid for new acquisitions and the cost of building things internally stay at the heart of the model.
Buying more companies also does not, on its own, explain how future earnings will be sustained. And it can leave important risks concentrated.
Imagine several subsidiaries that all depend on the same platform or sell to the same customer segment. One change hits their cash generation at the same moment it demands new investment from all of them. The holding company faces several requests for capital just as the cash to fund them is shrinking.
A holding company with financial room could help those subsidiaries respond faster than any of them could alone. One that had already committed the cash elsewhere would have fewer options. Diversification has to account for shared threats and simultaneous investment needs, not just separate revenue lines.
The promise to buy and hold forever describes an intention to keep owning the business. Sustaining or replacing the portfolio’s earnings still takes a continuing stream of decisions about what to buy, where to invest, and what to harvest. The length of the commitment does not make those decisions for you.
When the original assumptions stop holding
In a model where subsidiaries send cash up to the holding company, an operator’s investment proposal competes with every other use of that money. Central allocation can steer capital toward the best opportunities in the group. It also makes the subsidiary’s strategy depend on what the holding company is willing and able to fund.
A small subsidiary can have a healthy percentage margin and very little room in absolute dollars. One substantial rebuild, one unexpected problem, or one commercial push can eat the whole thing. Buying a business while it is profitable does not mean it can stay profitable every quarter through every change it needs to make.
A project often needs two things: funding, and acceptance of lower earnings for a while. Retained cash can pay for the work, but spending recognized as a current expense still reduces current profit. Some development costs can be capitalized and charged over time, which changes when the expense shows up, not how much cash leaves.[5]
So a standing margin requirement can block a project even when the holding company has the cash to pay for it.
Leonard made a related point in the same 2013 letter. He suspected some managers were running on “crude heuristics like ‘make 20% EBITA, and you can invest the rest’” and said plainly that he disliked the approach. He wanted managers to monitor the returns on their investments in organic growth. The question was what an additional dollar would earn, not how much spending fits under a fixed margin target.[4]
That approach still allows a clear no. A proposal can have a weak expected return, exceed what the team can execute, or lose to a better opportunity in another subsidiary. What matters is carrying the allocation decision all the way through to the operating plan. If an expansion depended on an investment that will not be funded, its growth target needs to be reconsidered. Keeping the target requires a credible revised plan for reaching it with the resources available.
Now suppose the holding company has already committed the subsidiary’s expected cash to debt payments, distributions, or the next acquisition, and the subsidiary turns out to need more investment than anyone planned. Saying no may be a reasonable response to the commitments already made.
Maybe the original assumptions missed a need that was visible at the time. Maybe they were sound and the market moved. Those are different verdicts on the acquisition, but both call for a fresh decision about the subsidiary.
The answer might be to fund the renewal, cut the project’s scope, find a cheaper way to deliver it, or shift to harvesting. Those decisions need to carry through to the operating plan, so the operator is accountable for an outcome the revised strategy can reasonably support.
Debt payments and other fixed obligations cannot simply be revised to accommodate a subsidiary’s new strategy. That loss of flexibility belongs in the decision about how to finance the acquisition in the first place.
The acquisition model needs room for conditions to change, and the operating agreement needs a way to respond when they do.
What does the holding company add?
While I evaluated roughly a dozen potential acquisitions for Orbit, I looked at retention, competitive exposure, the state of the product, growth prospects, and fit with our existing apps. None of them closed.
The recurring question was how the economics would change under us. What was the founder personally doing, and what would it cost to replace that if they left? Did the code need rebuilding? Which acquisition channels were actually working, what did they cost, and how much headroom was left? Which of the proposed channels had never been tested?
Our engineering depth and shared finance and HR support made some acquisitions more attractive under Orbit than they would have been on their own. But adding another app could also displace development already planned for an existing product. An acquisition might save hiring a separate team while still consuming capacity that had a better use.
A business can be a good fit for a holding company that has the capabilities it needs and a poor fit for one that does not. That should shape both the decision to buy and the price.
There is a real opportunity here beyond finance and administration. A small subsidiary often needs specialist help with design, product, SEO, marketing, content, or sales operations, without needing a full-time hire in every one of those disciplines.
Access to that help can be part of the reason to join a holding company at all. An operator can want full autonomy over product and customer decisions and still welcome the help.
Shared specialists could serve several related subsidiaries. A services business already in the portfolio might provide some of it. The savings would come from sharing scarce expertise, avoiding duplicate hiring, and reusing what the group already knows. An internal charge on its own saves nothing. The group still pays for the people doing the work.
Common ownership makes it easier to build long working relationships and have candid conversations about what a subsidiary needs. Accountability still needs clear commitments on scope, price, capacity, and delivery. The provider’s incentives should track the results it helps subsidiaries achieve. Operators should have a real choice about using the service and a way to push back on poor work.
Beacon makes this explicit in its pitch to founders: product and engineering help, sales and marketing support, and operating leadership alongside the capital.[6]
Hands-off ownership works best when the subsidiary already has the capabilities its strategy requires. Where it does not, autonomy is worth more when it comes with practical support.
Give each subsidiary a mandate that can change
The holding company and the operator need an explicit agreement about the strategy for the subsidiary. Where one subsidiary runs materially different products, the same conversation should happen product by product.
Start with retention, competitive position, the investment the business needs, and the opportunities in front of it. Include an honest view of what happens if nothing changes. From there, both sides can agree on the strategy the holding company is willing and able to fund.
I would distinguish five mandates:
| Mandate | The decision | What success means |
|---|---|---|
| Sustain | Invest in the product and commercial work needed to keep the business competitive and economically healthy. | Retention, reliable service, and cash generation stay healthy after necessary investment, and the competitive position holds. |
| Renew | Undertake a substantial, bounded change to restore or protect the business, with an agreed budget, milestones, and intended end state. | The change resolves the defined problem and leaves the business able to sustain or grow at acceptable returns. |
| Grow | Invest in an additional market, product, or commercial opportunity. | Evidence of demand and a credible path to profitable expansion justify further investment, and the cost of acquiring and serving customers holds up as volume increases. |
| Harvest | Limit discretionary reinvestment and realize the remaining cash flows while meeting customer obligations. | Deliver the agreed cash plan over the remaining life, with realistic expectations for flat or declining revenue and profit. |
| Consolidate or close | Combine activities or end a product or business when its prospective value no longer justifies the resources required. | Complete a responsible transition and release resources after integration or wind-down costs. |
These mandates describe intent. A subsidiary can sustain its core product while running a bounded growth experiment on the side. The spending and the expectations for each should be explicit.
Renew is a defined intervention above ordinary upkeep, like replacing technology that has become a serious constraint. It needs an endpoint. If major interventions keep being necessary just to keep the business standing, those costs belong in the assessment of its ongoing economics. The business may still be attractive after paying them. If not, the strategy has to change.
Sustain needs the most vigilance. A product can keep working and keep reporting acceptable margins while it quietly loses ground. Replacing customers who leave, improving distribution, and keeping pace with what customers expect are all part of sustaining a business.
Defer that work long enough and the gap gets too expensive to close. What began as a sustain mandate drifts into a decline that no longer justifies the investment needed to reverse it. Retention, the reasons behind lost deals, customer feedback, and the cost of winning new business can reveal that deterioration before it becomes obvious in reported profit.
For every mandate, the holding company and operator should agree on:
- Resources: the cash and team capacity committed, including any access to shared specialists.
- Authority: who can approve spending and hiring, within what limits.
- Financial expectations: any accepted reduction in near-term earnings and the trajectory expected afterward.
- Evidence and review points: what would justify further investment, a change of mandate, or stopping.
- Incentives: the results the operator will be rewarded for.
For example, a growth mandate to test a new marketing channel could carry a fixed budget and a review date. Further funding would depend on evidence against agreed targets for customer acquisition cost and retention, not on the number of sign-ups.
That makes the initial commitment clear without treating a promising idea as a fully funded growth plan.
An operator doing approved renewal work should not be rewarded for postponing it to protect this quarter’s margin. An operator harvesting a declining business should be measured against the agreed cash trajectory, decline included. In both cases, the financial reporting should separate the cost of approved work from unrelated underperformance.
When the evidence or the available capital changes, the mandate gets reviewed. The operator owes the holding company alternatives and disciplined execution. The holding company owes the operator clarity about the strategy it is funding and the results it expects.
Permanent ownership, revisable plans
At Orbit, evaluating acquisitions, building products, investing in sales and marketing, and closing products were all part of one job. That gave me a close view of how differently a profitable software business can need to be run as its circumstances change.
The holding company model can support that work. Capital can move toward the best opportunities. Operators can get access to capabilities they could never afford alone. Subsidiaries can be developed without a clock running on a resale.
The acquisition decision still matters enormously. The holding company should find out what sustaining the business will take, name what remains uncertain, and decide whether the opportunity fits its capabilities and its financial commitments.
After the deal, the holding company and the operator have to keep making that assessment together. A growth opportunity has to earn its funding. A sustain mandate has to keep pace with the market. Renewal needs a destination. Harvesting needs an honest cash plan.
Those choices are how a collection of acquisitions keeps earning worthwhile returns as its original products and assumptions change.
The commitment to ownership can be permanent. The operating mandate has to stay open to change.