Your tax credit practice captures what qualified companies have not yet claimed. Your pipeline captures the same three CPAs.

Your pipeline probably looks healthy on paper. A few CPA firms send you clients each quarter. A regional accounting partner recommends you when the R&D credit is too specialized for their team. A manufacturing client refers you to their trade group. Then the referrals thin out. The same three names appear on your call sheet. You wait.

This is the normal condition for tax credit capture firms. The problem is not a bad quarter. The problem is that your new business engine was built inside a network that has already mapped itself.

What the stall looks like

The symptoms are specific to this work. You know the credit inside a statute others miss. You have the engineers, the attorneys, the method to defend the position. Your close rate on qualified opportunities is high. The issue is that qualified opportunities arrive in batches, not flows.

A good year often traces back to one relationship. A partner at a mid-size CPA firm retires and their successor prefers an in-house group. A corporate controller you worked with in 2019 changes employers and the new company already has a credit provider. The pipeline does not drift downward. It drops.

You compensate with conference attendance, white papers, or a revised website. These activities feel like marketing. They rarely produce the named, engaged prospects you need. The credit has a deadline. The prospect must be identified, educated, and engaged before the filing window closes. A general reputation does not solve this timing problem.

Referral networks are closed geometries

Tax credit capture lives inside the accounting ecosystem. Your referral sources are specific: CPA partners, corporate tax directors who have worked with you before, CFOs who moved from one client to another and brought your name. Each relationship was built on trust earned through a successful engagement, defended under audit, or delivered with documentation that survived scrutiny.

This trust is the asset and the ceiling. A CPA who refers you once will refer you again when the client profile fits. They will not refer you to their competitor's clients. They will not refer you ten times a year because their own practice has a fixed shape. The network is dense with reciprocity. Referrals flow back and forth between known parties. It is efficient. It is also bounded.

The geometry is closed. Every new referral source you add requires the same years of trust-building. You are not expanding the network. You are replicating the same unit at a slower pace.

Why the ceiling does not move

You might believe the answer is more relationships. More CPA lunches, more bar association events, more LinkedIn connection requests to corporate tax managers. This extends the network at the edges. It does not change the structure.

Each new referral source takes eighteen to thirty-six months to mature. They must see you deliver once, preferably twice, before they risk their client relationship on your work. During that period, you are invisible to the prospects they serve. You are waiting in the antechamber of someone else's credibility.

Meanwhile, the buyer universe is larger than the referral network suggests. Thousands of companies qualify for credits they do not claim. R&D credits sit unfiled in software companies that never met a specialized provider. Historic rehabilitation credits go unclaimed by developers whose attorneys do not know the statute. WOTC credits are missed by HR departments that process payroll and never speak to a tax specialist.

These buyers are not in the network. They do not know to search for you. They do not know the problem exists.

The actual shape of the buyer universe

Your buyers are identifiable. They have titles: VP of Tax, Corporate Controller, Director of R&D, General Counsel, CFO at a portfolio company. They sit in companies of specific sizes, industries, and capital structures. A $40 million manufacturer with a five-person engineering team qualifies for R&D credits. A $200 million developer with two historic properties qualifies for rehabilitation credits. A $15 million restaurant group qualifies for WOTC.

These companies are not hiding. They file public registrations, report NAICS codes, disclose patent activity, announce facility expansions. Their qualification signals are visible to someone who knows what to look for.

The problem is access. They do not attend the same conferences. They do not read the same publications. They do not ask their CPA about credits the CPA does not know to mention. The referral network reaches the buyers who already know to ask. It does not reach the buyers who do not know the question.

What outbound correspondence changes

Correspondence means letters and emails written to a named person at a qualified company. The channels are Email Correspondence, Direct Mail, and Retargeting, with the phone as follow-up. Retargeting reinforces the letters and emails through paid digital placements targeted to the same buyer profiles.

This is not a replacement for referrals. It is a parallel geometry. Your name arrives on the desk of a VP of Tax who has never heard of your firm. The letter names a specific credit, a specific qualification signal visible in their public record, and a specific path to a preliminary assessment. The email that follows is sequenced to the calendar of the credit filing deadline.

The buyer who did not know to ask now has a reason to respond. The buyer whose CPA never mentioned the credit now has a second opinion in hand. The geometry shifts from waiting inside a closed network to identifying and engaging the qualified universe directly.

The referral pipeline continues. The CPA who trusts you still refers you. The correspondence pipeline adds prospects who were never inside that trust circle. The two systems run on different logic. One depends on relationship density. The other depends on identification accuracy and message precision.

What the work actually requires

Correspondence for tax credit capture is not a volume exercise. It is a qualification exercise. The list must be built from signals that predict credit eligibility: R&D expenditure patterns, hiring volumes, capital project announcements, patent filings, geographic location relative to opportunity zones or historic districts. The message must name the specific statute, the specific deadline, and the specific risk of non-filing.

The sequence is timed to the tax calendar. A first quarter letter about R&D credits for the current tax year. A third quarter follow-up about amended returns for prior years. A year-end note about credits that expire if not claimed. The correspondence earns attention because it is relevant to the recipient's actual situation, not because it is frequent.

The phone follow-up is not a pitch. It is a confirmation that the letter arrived, an offer to send the preliminary qualification checklist, and a scheduling of the assessment call. The retargeting placement keeps the firm's name visible to the buyer who opened the email but did not reply, reinforcing the correspondence without replacing it.

Who this does not suit

This mechanism does not fit every tax credit capture firm. It is not suited to firms that depend entirely on a single rainmaker who closes by personal presence and will not delegate to a correspondence sequence. The principal must be willing to let the initial contact be handled by written correspondence and a follow-up call from a team member.

It is not suited to firms with no defined buyer list. If you cannot describe the industry, size, and qualification signal of your ideal prospect, you cannot build the correspondence list. The work requires precision at the front end.

It is not suited to firms that lack capacity to absorb new volume. Correspondence produces a predictable flow of assessment calls. If your team is already at capacity and you have no process to onboard clients efficiently, the pipeline will back up and the close rate will suffer.

It is not suited to firms that treat every credit as a custom engagement with no standardized qualification or delivery process. Correspondence scales when the initial assessment is repeatable. If every engagement requires a bespoke research phase before you can quote, the economics of outbound do not hold.

The structural choice

You can continue to build the referral network one relationship at a time. This is a respectable path. It produces high-trust introductions and a close rate that reflects that trust. It also produces a pipeline with a fixed ceiling, visible in the repetition of the same source names and the dependency of each good year on a single relationship holding.

Or you can add a parallel geometry that reaches the qualified buyers who are not inside the network. This requires building the buyer list, writing the correspondence, sequencing it to the tax calendar, and following up with discipline. It requires accepting that the first contact with a new prospect will be written, not relational, and that the principal's time enters later in the process.

The choice is not between good and bad methods. It is between a closed geometry and an open one. The referral network is closed by design. Correspondence is open by design. For a tax credit capture firm with capacity and a defined buyer, the open geometry is the path through the ceiling.