Onboarding clients into a portal they will actually use
Most client portals die in the first fortnight. A practical two-week onboarding plan for agencies that earns adoption instead of announcing it.
· 5 min read · Client delivery
Somewhere in your agency's history there is a portal that died quietly. Someone spent a weekend configuring it. Someone else wrote a warm welcome email with a numbered list of benefits. The client logged in once, found an empty screen wearing somebody else's logo, and went back to email. The portal is still there, technically. Nobody has looked at it since March.
This is not a story about one bad tool. It is the default outcome of client onboarding for agencies, and it repeats because the failure has almost nothing to do with features. The portal did not lose on functionality. It lost in the first two weeks, before anyone had formed a habit, at the exact moment when the client was deciding whether this thing was for them or merely for you. If you want a portal your clients actually use, that fortnight is the whole game.
Why clients abandon portals
Start with an uncomfortable truth: your client already has a way of asking you for things. It is called email, and from where they sit it works fine. They type a sentence, press send, and something happens. When you introduce a portal, you are asking a marketing manager with two hundred unread messages to learn a new tool so that your delivery team can be better organised. That is a trade she never agreed to, and she can decline it silently, simply by continuing to email you. Most portals ask the client to change behaviour while the agency changes nothing, and clients notice.
The second failure is emptiness. The classic onboarding sequence goes: sign the contract, create the account, send the invite. The client clicks the link on day one, full of mild goodwill, and finds a dashboard that says there are no requests yet. Nothing about their retainer, nothing about the landing page you discussed on Tuesday's call, nothing with a date on it. They are being asked to do the filing before there is anything in the filing cabinet. The portal gave them nothing back, so they gave it nothing either.
The third failure is quieter. The client hired your agency. The screen says the name of a software company they have never heard of, in colours that belong to nobody. It feels like being handed off, the delivery equivalent of a call centre transfer. A portal that wears the wrong brand tells the client this relationship has a middleman now. A white-label portal that carries your logo and your colours tells them the opposite: this is still us, just tidier.
The first two weeks
Onboarding is not the invite email. It is everything you do before and just after it. The plan below is boring on purpose, because the dramatic version (big announcement, training call, laminated PDF guide) is the version that fails.
- Seed the workspace before anyone is invited. Load the live work in. The retainer deliverables for this month, the homepage copy that is with your designer, the two requests from last week's call, each one with an honest date the client can see. When the invite lands, the first click should open a page that is already about them.
- Let email keep working. Do not declare an email amnesty, because you will lose. Instead, give the workspace its own email address and let anything mailed to it become a request automatically. The client keeps their habit. You get the structure. Over a month or two the habit migrates on its own, because replies and updates start living where the work lives. There is a longer argument for this in stop running client work from your inbox.
- Make the first screen their own work, with dates. Not a feature tour, not a welcome video. The single most persuasive thing a client can see on day one is the status of the thing they were about to chase, with a date next to it. A visible landing date does more for adoption than any onboarding sequence, because it answers the question before it is asked.
- One login, not five. If the client needs one place for files, another for approvals, a third for status and a fourth for messages, the portal has already failed, whatever the tools cost. Everything about their work should live behind one login: the request list, the board, the deliverable reviews, the conversation. Every extra credential is a small tax on adoption, and clients do not pay taxes voluntarily.
What adoption actually looks like
Adoption is not a login metric. It is a Thursday afternoon where the client was about to email you the words "where is this up to", opened the portal instead, saw the landing date, and closed the tab without sending anything. Nothing happened, and that nothing is the product working.
You will know it is happening because the chasing emails thin out first. Then requests start arriving through the workspace address rather than to whichever account manager the client likes best. Then, some weeks in, a client will comment on a deliverable inside the portal without being asked to, and you can quietly consider the onboarding finished.
The pattern underneath all of this is simple. A portal is not adopted because it was announced well. It is adopted because, on the day the client happened to need something, it already had the answer. Earn that moment three or four times and the habit forms. Miss it, and no amount of training calls will save you.
One practical note on cost, since it shapes how freely you can invite people. If your portal charges per user, you will hesitate before inviting the client's finance director, and the portal quietly becomes internal software. Pricing per client workspace, with client users free, removes that hesitation entirely; the pricing should never be the reason a stakeholder stays outside the fence.
Where HoopoeHQ fits
HoopoeHQ was built inside a working agency, replacing a heavyweight ticket tool that clients refused to log into, so the first fortnight described above is the exact problem it was shaped by. Each client gets one strictly isolated workspace wearing your brand, requests can arrive by email, and every request carries a landing date the client can see. The 14-day trial includes white-label and needs no card, which means you can seed a real client workspace before you invite a single person.