The failure mode is partial adoption
Most writing about portal rollouts treats the risk as rejection — you buy the thing, announce it, the clients refuse. That is not the failure that actually happens, and it is not the one worth designing against.
What happens is partial adoption. Half your clients start using the portal. The other half carry on emailing, or texting photographs, or handing you a carrier bag in March. Nobody has said no, nothing has visibly gone wrong, and you are now running two document channels instead of one.
Two channels is not one and a half times the work, because the cost is not the second inbox — it is the mental map. Before you can chase anybody you have to remember which channel they are on. At month end you are reconciling a portal dashboard against an email folder and holding the difference in your head. That is worse than where you started, where at least everything was in one untidy place.
The point of a rollout plan
Everything below is reasoning from that point rather than reporting from a study. The published evidence on portal adoption in small practices is thin to the point of being useless, and most of the numbers in circulation trace back to nothing checkable — the longer account of what anyone actually knows sets that out. Treat the plan here as a cheap way of getting your own evidence from your own client base.
You chose it. Your client did not.
There is a structural point underneath the two-channel problem, and it explains why rollouts drift rather than fail cleanly. The person who buys the portal and the person who has to use it are two different people, and only one of them wanted it.
You evaluated it. You read the pricing pages, weighed up whether you could get away with one you already own, and decided. Every part of that was yours. Your client did none of it. They are a builder or a caterer or a physiotherapist who currently photographs a receipt and texts it to you, which works perfectly from where they are standing. They have no relationship with the vendor and no interest in the vendor. They get no benefit from switching — the benefit accrues to you — and they are the one being asked to change a habit.
It is also unlikely they have met a portal from anyone else. In the Wolters Kluwer Annual Accounting Industry Survey, of 2,373 respondents, 17% said they planned to invest in a client portal. That is a minority intention across the profession, which means the thing you are introducing is probably not something your client has been asked to do by their previous accountant, their solicitor or their insurance broker. You are asking for a new behaviour, not a familiar one.
Why the big-firm playbook does not port
A thirty-person firm can mandate a portal across its client base. It has the authority, an onboarding process that runs the same way every time, someone whose job includes chasing adoption, and clients who expect formality from a firm that size. It can write the portal into the engagement letter and enforce it. A solo practitioner asking a long-standing client to change habits has none of that, and no leverage at all. Pretending otherwise is how rollouts go wrong. Every step below assumes you are persuading one person at a time, because that is the actual situation.
Pick the month before you pick the clients
This gets decided last and should be decided first. A new tool that is not yet embedded is exactly what gets dropped the moment pressure arrives — by you as much as the client.
- Not January. If you work with US clients, busy season runs late January to 15 April and nothing introduced in that window survives it. The UK equivalent is the run-up to 31 January. Find your own jurisdiction’s crush and stay out of it.
- Not the month before a statutory deadline. Quarter-end VAT, BAS, payroll year end. Attention is the scarce resource and a deadline consumes all of it.
- A quiet month, with a second quiet month behind it. Two full cycles is the minimum that teaches you anything, so you need two quiet months in a row.
If you cannot find two consecutive quiet months, that is worth knowing before you spend money. It may mean staging this across a longer period, or fixing the close first — the month-end close checklist is the cheaper intervention if what you have is a process problem rather than a channel problem.
Start with six clients, not sixty
Six is small enough that you can hold the whole thing in your head, and large enough that one enthusiast and one refusenik do not decide the result. The number matters less than the order of magnitude: single figures, not the whole book.
Pick the six most responsive, not the six most annoying
The instinct is to start with the client who never sends anything, because they are the reason you bought the portal. That is the wrong first move. At this stage you are not testing whether difficult clients will adopt — you are testing whether your own process works — whether your request wording is clear, whether your item list is right, whether the reminders land, whether the files come back in a state you can use. You cannot learn any of that while simultaneously fighting an adoption battle. Two variables, one experiment, no usable answer.
So pick the six who already send things on time. If your process is broken, cooperative clients will tell you clearly, quickly, and without it costing you a relationship. The client who never sends receipts is a later problem and possibly a different problem altogether.
What to say to the first six
Send these individually. Not as a blast, not as a firm announcement, not from a no-reply address. One at a time, from you, to a person whose name you use.
The introduction
Short. Framed as fewer emails for them, not as better security for you. Names exactly what to do next, in one step. Does not explain the software.
Three things that email deliberately does not do. It does not use the word portal, which sounds like a system to log into. It does not say “more secure”, for reasons below. And it does not ask permission, because a question invites a negotiation you do not want to have; it states what is happening and offers a way out.
The first-month nudge
Some of the six will not open the link. Assume nothing about why — it went to spam, they meant to do it on Sunday, they opened it on a phone in a car park and gave up. A nudge that assumes reluctance creates reluctance where there was none.
The out matters. You are gathering information, and a client who quietly ignores the link for three months tells you nothing, while a client who replies “I would rather email” has just given you a data point and saved you a quarter of chasing.
The three things to record per client
Write these down; a sheet with six rows is enough. Writing it rather than remembering it is the same principle that makes two channels expensive: the mental map is what costs you.
| What you record | What it tells you | What it does not tell you |
|---|---|---|
| Did they open the link? | Whether the email got through and was legible. A client who never opens has a delivery or attention problem, not an adoption problem. | Whether they will use it. Opening is not using. |
| Did they upload at least one file? | Whether the page is usable by an ordinary person on whatever device they own. | Whether it stuck. A first upload is often just politeness. |
| Did they upload again the following month? | Everything. The only number that separates a habit from a favour, and the one the decision rests on. | Nothing you need at this stage. |
The third row is the one people skip and the only one that matters. A client will use a new tool once because you asked. Whether they come back and do it again unprompted the next month is the difference between a channel and a novelty.
Run one full monthly cycle, not a fortnight
The 60-day test
After two cycles you need a decision rule you set in advance, because a rule you invent afterwards will be shaped by how much you have already spent.
The rule I would use: if fewer than roughly 60% of the clients you rolled out to are uploading again in month two, the tool is creating work rather than removing it for this client base, and no amount of better reminding will fix that.
Where 60% comes from
It is a judgement and a rule of thumb, not a finding. There is no industry benchmark behind it and I am not going to dress one up. The reasoning is only this: below about six in ten, the minority still on email is large enough that you are permanently running two channels, and the portal has become an additional place to look rather than the place to look. Pick your own threshold if you prefer — but pick it before you start and write it down. The point is that the number exists in advance, not that it is 60.
Above the line, expand. Below it, with the six most cooperative clients you have, be honest that the rest of your book will do worse rather than better. The real failure would be discovering that at client forty.
Write the opt-out policy before you need it
You need a written opt-out policy before you send the first link, not after a client refuses. Written mid-refusal it becomes a negotiation with one person; written in advance it is a rule applied consistently, which is far easier to say out loud and to hold to. There are three honest options and there is not a fourth.
- Keep them on email permanently, and accept the two-channel cost for a named minority. This is defensible as long as the list is written down, short, and closed. “Three clients are on email” is a manageable exception. “Some clients are on email” is the mental map again, and it will cost you.
- Price the extra handling into their fee. Collecting documents by email and reconciling them by hand is work. If it is permanent for a particular client, it can sit in that client’s fee rather than in your evenings — a conversation for the next fee review, not the same breath as the portal.
- Accept that they may not be a fit. The far end of the scale, and rare. Worth noting that a vendor in this space publishes advice along the same lines — that a client who will not work remotely may be better served by a local specialist. When portal vendors themselves acknowledge some clients are not convertible, you are entitled to conclude the same about one or two of yours.
The conversation with a client who has refused
Do not fight it, do not re-sell the portal, and do not make the client wrong. They are allowed to prefer email. What you are doing here is closing the question so it stops consuming attention every month.
Then add them to the written list of email clients, and stop thinking about it. The cost of an opt-out you have decided on is much lower than the cost of one you keep reopening.
Expanding in batches, worst cases last
Assuming you passed the 60-day test, expand in batches rather than in one movement, and keep the order deliberate.
- Batch two: the next tier of responsive clients. Ten or twelve. You are testing volume rather than mechanics now — whether your routine holds with eighteen open requests instead of six.
- Batch three: the bulk of the book. The wording is settled by now and you know which questions get asked.
- Last: the difficult ones. The client who has not sent a receipt since March goes at the end, not the beginning. By then you know the process works for everyone else, so when it fails with them you can be confident it is about them and not the tool — which is the input to the fee conversation.
At every stage, keep the written list of who is on which channel current. One column, two values, a minute a month. It is the single thing that stops partial adoption becoming the expensive version of itself.
Set a decision date for each batch
Two cycles per batch, then a decision: portal, or email. What you must not do is leave a batch unresolved and start the next one. Three unresolved batches is the two-channel problem with extra steps, and it is where most rollouts end up.
What not to do
- Do not announce it as a firm-wide change by email blast. A blast invites a firm-wide referendum, and you will lose it — not because the portal is bad but because a group asked to approve a change to their habits will decline. Individual emails get individual answers, and individual answers are mostly yes.
- Do not lead with security. It is true that a portal handles documents better than email attachments, and your client does not care. Security is a benefit to you and to a regulator, not to the person holding the phone. Lead with fewer emails and one place to look, which are benefits they can feel.
- Do not run both channels indefinitely without a decision date. This is the whole article in one line. Both channels for a defined pilot is a plan; both channels with no end is the failure mode.
- Do not roll out into busy season or a statutory deadline. Covered above, and worth repeating because it is the most common way a sensible plan dies. A tool that is not yet embedded is the first thing dropped when the pressure arrives.
One note on the software, since this is our site. ClientVault fits this shape of job: the client gets a link and uploads with no account and no password, follow-up emails go out automatically on day 3, day 7 and day 14 after a request is created, recurring monthly, quarterly or annual requests send themselves, and there is a free tier for three clients — about the size of a first pilot. What it does not do is solve the adoption problem. Nothing does. The staged rollout above is the work, and it is the same work whichever tool you pick, including the free one you may already have. Skip the staging and a better upload page will not save you from the two-channel trap; it will just be where half the documents are.