Why Do Half Your Partners Ignore the New System?

A firm of forty people with six partners spends most of a year selecting a practice management platform. The demos are impressive. Implementation goes reasonably well. Templates get built, everybody is trained, and the launch is announced.

Eighteen months later, two partners run their entire book inside it. Two use it for some things and keep their own arrangements for the rest. Two have quietly gone back to what they were doing before, which is a spreadsheet and a very good memory.

The managing partner now has the worst of both situations. The firm is paying per person per month for software that describes only part of the work, and the reports it produces are confidently wrong, because they reflect the partners who complied rather than the firm. So nobody trusts the reports, which means nobody uses them, which confirms the sceptics.

That is the mid-market problem with practice management automation for accounting firms, and it is almost never described honestly, because describing it correctly means admitting that the software was not the answer.

Here is what matters most:

• Platforms are built around one process. A firm where partners each own a book of business has several, and they are different for defensible reasons.
• Partial adoption is worse than none, because the reporting becomes unreliable while the cost stays fixed.
• Two vendors publishing guides in this space state in their own words that they offer no specific guidance for 25 to 75 person multi-partner firms.
• Even the profession's own benchmark survey is dominated by firms under five million dollars of revenue, so mid-market operating data is genuinely scarce.
• The fix is not more training or another platform. It is automation shaped around how each partner actually runs their book.

Your partners are not being difficult

This is the part vendors and consultants get wrong, and getting it wrong is why change programmes fail.

When a partner declines to work the way the new system expects, the standard reading is resistance, and the standard prescription is executive sponsorship and more training. Occasionally that is correct. Far more often the partner is protecting something real.

Consider what differs legitimately between two partners at the same firm. One runs two hundred individual returns with a heavy seasonal peak and needs to know only whether documents have arrived. The other runs thirty complex entity groups where a single client's work spans four deliverables and three staff over eight months, and needs to know which stage each deliverable has reached. A workflow designed for the first partner is useless to the second, and one designed for the second is oppressive overhead for the first.

Now add the arrangements that exist in every established firm and appear in no software design. A partner who reviews everything personally and one who delegates review to a senior manager. A partner whose clients contact them directly by phone and one who insists everything comes through the firm. Two offices with different administrative staffing. An inherited book from a retired partner that was never reorganised because the clients are elderly and the fees are good.

None of that is inefficiency to be trained out. It is the accumulated shape of a real practice. A platform that requires one process asks six partners to abandon their own judgement about how to serve their own clients, and the two who refuse are often the two with the most valuable books.

Nobody is writing for a firm your size

Look for guidance on this specifically and you will notice something odd. The material divides cleanly into content for very small firms, where the answer is a single sensible workflow, and content for large firms, where a dedicated operations function absorbs the complexity. The band in between is thin.

This is not my impression. Two vendors publishing comparison guides in this category say it in their own text. One notes that its guidance for 25 to 75 person multi-partner firms is minimal. Another states that multi-partner and multi-office structures receive little specific attention in its guide and are largely treated as scaled versions of single-office practices. That phrase is worth sitting with, because it is exactly the design assumption that fails: a forty-person firm with six partners is not a bigger version of a six-person firm. It is a federation.

The data gap is real too. The profession's own benchmark study, the AICPA and CIMA National MAP Survey, gathered over a thousand responses on the last full year. It found median net client fees growing 6.7% year on year, down from 9.1% in the prior study, and net remaining income per partner rising 11.9% across two years. Useful figures. But roughly four in five respondents are firms under five million dollars of revenue, which means even the best independent data available under-represents the firm shape being discussed here. When partners tell me there is no good benchmark for a firm like theirs, they are right.

What partial adoption actually costs

Three costs, and firms usually notice the smallest.

The visible one is licence spend on capability nobody uses. Practice management sits in the region of fifty to eighty dollars per person per month at most firms, charged for everybody regardless of whether that person's work is described in the system. That is the cost partners raise in meetings and it is the least important of the three.

The second is the reporting. This is the one that quietly does damage. A capacity report that covers four partners out of six is not partially useful, it is misleading, because the gaps are invisible in the output. Decisions get made on it: which work to accept, who needs help in March, whether to hire. A managing partner making resourcing calls from a system describing two thirds of the firm is worse off than one making them from experience, because the number carries an authority it has not earned.

The third is what happens to the partners who complied. They have taken on administrative overhead to feed a system whose outputs the firm does not trust, while their colleagues carry none. That is precisely how a good initiative acquires a bad reputation, and once it has one, the next attempt is harder regardless of merit.

What actually works

The change in thinking is straightforward. Stop asking the partners to converge on the software's process. Let the automation adapt to each partner's process, and standardise the small number of things the firm genuinely needs to be identical.

In practice there are usually only three or four of those. That a client's work is visible somewhere. That a deliverable's status means the same thing across the firm. That completed work becomes an invoice. That nothing sits untouched past a threshold nobody has to remember. Everything else can differ by partner without costing the firm anything at all.

What that looks like from a partner's chair: the way you run your book does not change. The information the firm needs appears without you assembling it, because it is derived from things that already happen rather than from you filling in a system. The partner who only cares whether documents have arrived sees exactly that. The partner tracking four deliverables across eight months sees those. The managing partner sees a firm-wide position that is accurate because it did not depend on six people complying with a process none of them designed.

This is also why it cannot be bought. A platform vendor cannot ship this, because the configuration is your firm's partner arrangements, and those are not a product feature. It is the same reason your practice management implementation, however well run, stopped at configuring the software: the published scope of that work typically excludes customising it to your specific clients and internal processes. That part was always going to be yours.

It usually pairs with the two places mid-market firms lose the most time, which are the document-to-preparation gap and accepting a new client. All three have the same underlying cause: each system holds one part of the work, and your people are the connection between them.

The objections that deserve answers

Are we just institutionalising bad habits? Sometimes, and it is a fair challenge. The honest distinction is between variation that serves clients and variation that is merely historical. A partner who reviews personally because their clients are complex is different from a partner who reviews personally because nobody has revisited it since 2014. The exercise of writing down what each partner actually does surfaces that difference quickly, and it is usually the partners themselves who volunteer which of their habits are no longer worth defending.

Should we not simply enforce one process? You can, and some firms do it successfully, but be clear-eyed about the price. It works where the managing partner has genuine authority over how partners run their books, which is uncommon in a partnership, and where the client work is homogeneous enough that one process fits. Where it fails, it fails in the manner described at the top of this article, and you will have spent the political capital for a system that describes two thirds of the firm.

Does this scale if we merge or add partners? Better than a single enforced process, in our experience, because a merger is precisely the moment a firm acquires partners with established ways of working and no appetite for abandoning them in year one. A structure that already accommodates variation absorbs an incoming group without a change programme.

FAQ

How do we know if we have this problem?

Ask two questions. First, request the same capacity or work-in-progress figure from your system and from each partner independently, and compare. If they disagree materially, your reporting is describing part of the firm. Second, ask each partner to describe how they know where a client's work stands. If the answers differ in kind rather than in detail, one process was never going to fit, and no amount of training was going to change that.

Do we need to replace our practice management platform?

Almost certainly not, and replacing it is the most expensive answer available. It costs the implementation you already paid for, the templates already built, the training already absorbed, and a season of everyone being slower, and it delivers you to the same problem with different branding, because the next platform will also assume one process. The cheaper path is to keep the platform as the place the firm looks and build the connective work around it.

How long before the managing partner can trust the numbers?

The realistic answer is one full cycle of whatever you are measuring, because trust comes from a partner checking the number against what they know and finding it right. That means starting with one measure that matters, getting it accurate for every partner including the two who ignore the system, and letting people test it. Firms that try to produce a complete dashboard first usually produce something nobody checks, and an unchecked number is indistinguishable from a wrong one.

Where does this sit against hiring an operations manager?

They solve different halves and the good version is usually both. A capable operations person is worth hiring at this size and will immediately be more effective if the firm's information is reliable, rather than spending their first year manually reconciling six partners' arrangements into one report. Hiring someone to be the connection between your systems works, but it is an expensive way to be a connector, and it leaves with them.

Read next: AI Automation for CPA & Accounting Firms