Build vs buy: in-house origination or a partner.

Build in-house when origination is a permanent function attached to a thesis you expect to still be running in five years, and you can fund a dedicated seat plus the data stack that sits under it. Buy a partner when you need coverage of a defined mandate faster than you can hire for it, when the thesis may change inside a year, or when the outreach volume required does not justify a full-time salary.

Most funds treat this as a two-option decision. It is a three-option decision, and the third option is the one that quietly absorbs the budget: buying sourcing software is not buying origination. Software gives you a list. Someone still has to work it.

This page is written for the person actually making the call: a partner, a head of origination, an independent sponsor or a search fund principal who has a mandate, a limited amount of time, and a board or LP base that will ask why the pipeline looks the way it does.

01What are the three real options?

The honest version of the comparison is that these are not mutually exclusive, and the funds that get this right usually end up with a deliberate mix rather than a clean choice. The failure mode is arriving at the mix by accident: a licence renewed out of habit, a junior hire absorbed into diligence, and no one accountable for outbound coverage.

02What does each option actually cost?

Take the software line first, because it is the one that appears in all three columns.

Sourcing platforms in this category do not publish prices. Grata lists three tiers, Growth, Scale and Alpha, plus add-ons, with no figures attached and a "talk to a specialist" route to a quote (grata.com/pricing). Third-party buyer data aggregator CostBench puts the median Grata contract at around $155,000 per year, based on six reported buyer contracts (CostBench); treat that as an indicative third-party figure rather than a quoted price, because your number will depend on seats, data scope and term.

One structural change worth pricing in: Grata and SourceScrub are combining, announced on Grata's own site (grata.com/pricing). Two of the three obvious independent alternatives in this category are becoming one vendor. If your build case assumes competitive tension at renewal, that assumption is weaker than it was.

Now the salary line. Published private equity compensation data is heavily weighted to mega-funds and is not a clean proxy for a lower-middle-market sourcing seat. Wall Street Prep puts a first-year associate at $135k to $155k base with a $140k to $230k bonus, and a second-year associate at $160k to $180k base with a $170k to $270k bonus, noting the data came from a VP at a global alternative investment manager and that lower and middle market firms sit at the bottom of those ranges (Wall Street Prep). That is the ceiling, not the number you should plan against.

So build the number yourself rather than lifting a published range, and build it up honestly. A fully loaded in-house origination seat is not the salary. It is salary, bonus, employer costs, the data subscription the person needs to do the job, a CRM seat, and the partner time spent managing, reviewing and coaching them. The data line does not disappear when you hire; it moves into the same budget as the person.

And the partner line. An origination engagement is a fee for coverage of a defined mandate, not a salary and not a licence, which is why it is the only one of the three that scales down at the end of a term. What moves it is the size of the universe, the number of sectors and geographies in scope, and how much research each approach needs before it is worth sending.

A fuller cost breakdown, including the variables that move each of these figures, sits on our separate page on what origination actually costs.

03How long until each option produces a qualified conversation?

Compare ramp honestly, because it is where in-house cases are usually overstated.

The in-house clock does not start when the analyst signs. It starts when you open the search. Hiring cycle, notice period, onboarding, learning your thesis, learning how you talk to owners, first outreach, first replies. Every one of those is real elapsed time, and the first two are entirely outside your control.

Software ramp is fast on paper and slow in practice. Access is immediate. The behaviour change is not. A licence bought in March and first genuinely worked in September has cost you six months of coverage and a full year of fee.

A partner's ramp is onboarding to the mandate: agreeing the thesis, the size band, the geography, the disqualifiers, and the message. That is faster than hiring, but it is not zero, and any partner who tells you they will produce owner conversations in week one is selling you something they cannot deliver.

Whatever number a partner gives you, check what it is measured from. A first qualified owner conversation comes out of a cycle of research, approach and follow-up rather than a single send, so measure from the start of the mandate to the first real conversation, not to the first email sent.

For our own numbers: once sending starts, a new campaign on our PE and M&A mandates gets its first positive owner reply a median of about a day and a half after its first email. What varies is the onboarding before that. On our documented healthcare investment bank engagement, the mandate produced 14 qualified founder conversations in its first three weeks and 133 within 90 days. The full data is in our deal origination benchmarks.

04Which model gives better market coverage?

Coverage is the argument for buying, and it is the strongest one.

With Intelligence's Deal Origination Benchmark Report 2026, which aggregates pipeline data from 185 private equity firms across eight peer groups and only includes firms reporting consistently for at least two years, finds a significant contraction in median target market size across every peer group, and describes the dispersion in market coverage and sourcing effectiveness between firms as pronounced (With Intelligence). The same report notes that fewer than half of participating firms completed a platform exit in 2025.

Read that together and the implication is uncomfortable but useful: the addressable universe for a given thesis is smaller than it was, and the gap between the funds who cover their market well and the funds who do not is wide. Coverage is now a differentiator rather than a hygiene factor.

One person, however good, covers what one person can cover. That is the ceiling on the build option, and it is a hard ceiling. If your thesis has a universe of a few hundred companies, one dedicated analyst can cover it properly and build real relationships over time. If the universe runs to several thousand across multiple sectors or geographies, a single seat cannot touch it at a meaningful frequency, and you will end up with a pipeline shaped by whatever the analyst had time for rather than by your thesis.

05What about confidentiality and conflicts?

This is where most build cases are actually decided, and pretending otherwise is dishonest.

An in-house team is structurally cleaner. Your thesis stays inside your walls, your approach to a specific owner is yours alone, and there is no question about who else is being told what you are looking for.

An outsourced partner has to answer three questions in writing, not in a sales conversation:

1. Do you take another mandate in my sector and size band at the same time, and if so, what actually separates them?
2. Who sees my thesis, my target list and my notes?
3. What happens to the list and the research when the engagement ends?

If a partner cannot answer all three plainly, that is your answer. Ask us the same three and we will answer them in writing before you sign anything. A policy that exists only in a sales conversation is not a policy.

There is a second confidentiality point that cuts the other way, and it is rarely raised. Early-stage approaches to owners in a tight sector travel. If a fund's own name is on every first approach, the market learns exactly what that fund is hunting, and competitors read the same signal. An intermediated first approach keeps the thesis quieter for longer. Whether that matters depends on how distinctive your thesis is.

06What breaks in an in-house origination function?

Four things, in roughly this order of frequency.

The analyst gets absorbed. A live deal appears, and the person hired to source is pulled into diligence, modelling and IC prep, because that work has a deadline and outbound does not. Outreach stops. Nobody notices for a quarter. This is the single most common failure, and it is a structural problem, not a discipline problem: origination is the only function in a fund with no external deadline.

The role is a stepping stone. Sourcing is widely treated as a route into deal execution. Good people do it for a while and then want the next thing. When they leave, the relationships, the context on why a given owner said not yet, and the reason a company was disqualified two years ago all leave with them, unless it was written down somewhere other than their inbox.

The thesis moves. A permanent hire is a fixed cost against a thesis that may not survive the next fund cycle or the next macro turn. You cannot scale a salary down for two quarters.

Volume is underestimated. The arithmetic of outbound origination is unforgiving. Reaching a small number of genuinely qualified owner conversations requires a much larger number of researched, personalised, correctly-timed approaches, plus persistent follow-up over months. Funds consistently budget for the research and underbudget for the follow-up.

07What breaks with an outsourced partner?

Three things, and they are just as real.

A vague mandate produces vague output. "Good businesses in industrials, $5m to $20m EBITDA" is not a mandate, it is a category. A partner working from a vague brief will bring you companies that technically qualify and that you would never buy, and you will conclude that outsourcing does not work when the actual failure was the brief.

No capacity to take the conversations. A partner that succeeds creates owner conversations. If nobody on your side can take a call within a few days of an owner raising their hand, that owner cools, and the effort is wasted. Origination fails on your side of the line as often as it fails on the partner's side.

Expecting deals rather than conversations. An origination function produces qualified conversations with owners. It does not produce closed deals, and any partner who prices or promises on closed deals is either taking a success fee that changes their incentives, or overselling.

We have written separately about who we are not a fit for, and about what a realistic first 90 days looks like.

08Who should genuinely build in-house?

Build if most of these are true:

• Origination is a permanent function for you, not a project attached to one fund or one thesis.
• Your target universe is small enough that one dedicated person can cover it at a meaningful frequency and build real relationships over years.
• You can protect the seat from being absorbed into deal execution, and you are willing to enforce that when a live deal appears.
• Someone senior will actually manage the role, review the message, and coach it. An unmanaged sourcing analyst produces activity, not pipeline.
• Confidentiality on your thesis is worth paying a premium for, and you have concluded no external party should see it.
• You can absorb the gap when that person leaves.

Build is the right answer more often than partners like us admit. If you have a durable thesis, a defined universe and a partner willing to manage the seat, an in-house origination function compounds in a way that a bought one does not.

09Who should buy?

Buy if most of these are true:

• The mandate is defined and time-bound, and you need coverage now rather than after a hiring cycle.
• Your target universe is larger than one person can cover properly.
• The thesis may change, and you do not want a permanent salary attached to a temporary view.
• The senior people who would otherwise do the sourcing are worth more doing something else, and are currently doing outreach badly and irregularly.
• You want the function to exist without also building the management overhead around it.
• You want to test whether systematic outbound origination works for your mandate before committing to a permanent seat.

10Should you do both?

Frequently, yes, and the sequencing matters.

The common pattern that works: buy first to prove the mandate is workable and to learn what messaging actually gets owner replies in your sector, then hire in-house once you know the shape of the job and can write a job description based on evidence rather than a guess. The hire ramps faster because there is a documented playbook to inherit.

The pattern that does not work: hire first, watch outbound stall inside two quarters, then bring in a partner to fix it while the salary is still running. You are now paying twice for a function that has already lost credibility internally.

If you do run both, split by coverage rather than by task. Give the in-house seat the accounts where relationship depth compounds, the ones you will still be talking to in three years, and give the partner breadth across the rest of the universe. Splitting by task, where the partner researches and the in-house person sends, is where handover friction eats the benefit of both.

11How to decide in one sitting

Answer these six, honestly, and the decision usually makes itself.

1. How many companies are actually in your target universe? If the answer is a few hundred, one person can cover it. If it is thousands, one person cannot.
2. Will this thesis still be live in three years? If not, do not attach a permanent salary to it.
3. Who protects the sourcing seat when a live deal lands? Name the person. If you cannot, the seat will be absorbed.
4. What is your fully loaded cost per qualified owner conversation under each model? Not cost per name, cost per conversation.
5. Can you take an owner call within 48 hours of them raising their hand? If not, fix that before you spend anything on either option.
6. What happens to the pipeline knowledge if the person, or the partner, goes away next quarter?