The first part of this series looked at what the data says: roughly 250 European unicorns have produced more than 1,400 startups founded by their former employees, according to research from Accel and Dealroom. The pattern is real, it is concentrated in a small number of companies, and it clusters geographically.
Knowing that is not the same as being able to act on it. A corporate investor who reads the alumni data and concludes "we should track unicorn alumni" has identified a signal, not built a capability. This piece is about the distance between the two, because that distance is where most corporate sourcing initiatives quietly fail.
The Signal Has a Shelf Life
The single most important property of the alumni signal is that it decays. A map of who left which company, when, and what they are building now is accurate for roughly a quarter. After that it is a historical document.
This is what separates alumni tracking from most other sourcing inputs. A sector thesis stays broadly valid for a year. A relationship with a fund stays valid for as long as the relationship does. A list of people who recently left a scaling company is perishable in a way that neither of those are, because the entire value of the signal is that it identifies founders before they are visible, and visibility arrives on its own within months.
That property determines everything about how the capability has to be built. A one-off mapping exercise, however thorough, produces something that is wrong within two quarters. The question is never whether to build the map. It is whether the organisation is prepared to fund the maintenance.
What the Map Actually Needs to Contain
Three fields do most of the work, and they are less obvious than the ones most teams start with.
The departure, not the arrival. Most tracking starts when a new company is registered or announces a round. By then the signal has already been consumed by everyone else. The useful entry is the departure of a senior operator from a company that fits the thesis, recorded when it happens, with no assumption about what they will do next.
The seniority and the function. Not every departure carries the same weight. The alumni research shows the effect is driven by people who held operating responsibility, not by headcount generally. A departing head of product from a payments company is a different signal from a departing sales hire.
The gap. The time between a departure and the first visible activity is itself informative. A founder who surfaces within weeks was probably building before leaving. One who surfaces after a year has usually taken a genuine decision, and is more likely to be reachable during that gap, which is precisely when a corporate investor can be useful and nobody else is calling.
Where Corporates Get the Geography Wrong
The research finding that more than half of alumni-founded startups launch in the same city as their originating company is the most operationally useful thing in the dataset, and the most frequently ignored.
It means the map does not need to be global to be valuable. A corporate whose strategic interest is European insurance distribution does not need to track alumni of every unicorn in Europe. It needs to track a small number of companies in a small number of cities, which is a tractable problem for one person, and an intractable one for a team that insists on comprehensive coverage first.
The counterweight is that founders relocate, and the mobility of European deep tech founders is a real pattern of its own. A map built only on the origin city will lose the founders who leave, which is why the map records the person rather than the address.
The Part That Decides Whether It Works
Most corporates that attempt this get through the build and fail at the maintenance, for a reason that has nothing to do with the quality of the initial work.
Building the map is a project. It has a scope, an end date, and someone visibly responsible for it. Maintaining it is a job. It has no end date, produces nothing visible in any given week, and is the first thing dropped when the person holding it has a busier quarter than usual.
Organisations fund projects easily and jobs reluctantly. This is why the useful question, before any of the methodology, is which named person owns the weekly update and what else that person will stop doing. A corporate that cannot answer that question will build a map, use it well for one quarter, and quietly stop.
What This Buys That Intermediated Sourcing Does Not
A pipeline assembled from bankers, demo days and inbound decks can be reproduced by any competitor at the same price. That is not a criticism of it, it is a description of what it is: a subscription to the same information everyone else has, priced accordingly in every contested round.
The alumni map is different in one specific way. It is built from the corporate's own decision about which companies matter in its sector, maintained with its own effort, and therefore not available to a competitor who has not made the same decision and done the same work. Whether that is worth the maintenance cost depends on how many decisions per year the corporate expects to make on the back of it, which is the same variable that governs where capital is concentrating and how any venture structure should be sized.
How to Start Without Building Everything
The version of this that works starts far smaller than most teams expect.
Pick five to eight companies whose alumni would plausibly build something the corporate would want to see first. Record senior departures from those companies as they happen, with the date and the function. Review the list once a month, and reach out to the people who have not yet surfaced publicly. Nothing in that requires a tool, a data licence, or a headcount.
The discipline it does require is the willingness to keep doing it in the months where nothing comes of it, which is most months. The value arrives unevenly, and a programme that judges the effort on any single quarter will conclude, correctly for that quarter and wrongly overall, that it is not working.
Frequently Asked Questions
How large should the tracked company list be?
Small enough that one person can review it monthly without it becoming someone's full-time job. In practice that means single digits, chosen against the thesis rather than by size.
Is this a replacement for working with funds?
No. It answers a different question. Funds give breadth and pricing discipline; the alumni map gives early access to a narrow set of founders in a chosen sector. Most corporates need both, for different reasons.
What if the founders are not raising yet?
That is the point. The window where a corporate is genuinely useful, and genuinely differentiated, is before a round exists. A conversation at that stage costs nothing and is remembered.
Can this be outsourced?
The mapping can. The judgement about which companies belong on the list, and the relationships that follow from it, cannot — those are the parts that make the signal proprietary in the first place.
Conclusion
The alumni data is genuinely predictive, and it is available to anyone who reads the research. What is not available to anyone is a maintained map of a chosen set of companies, built on a corporate's own view of its sector, updated by someone whose job it is.
The gap between the signal and the capability is not analytical. It is organisational, and it comes down to whether a named person owns the update. Everything else in this piece is detail.
Working on this? Mandalore Partners runs corporate venture programmes for insurers and financial institutions, including the sourcing layer described above. Talk to us.