How to get client approval on selections without chasing email

A working approval process for interior design selections: what belongs on every item, why three answers beat two, what to do with a decline, and how to keep a record that survives a dispute.

By Idan Mann, founder of NYA

Most approval processes fail in the same place. The designer sends a beautiful PDF, the client replies to the email with "love 1, 3 and 4, not sure about the sconce, what does the rug cost again?", and the project's decisions are now scattered across a thread that nobody can sort, count or produce later. The presentation was never the problem. The reply channel was.

Approve the item, not the document

A twelve-page selections PDF asks for one answer to twelve questions, so you get one vague answer. Break the same content into individual items and each one carries its own state, its own date and its own answer. You stop asking "what did they think of the presentation?" and start seeing that nine items are approved, two are pending and one is declined — which is a schedule, not a feeling.

It also changes what happens when one thing is wrong. Under a document-level approval, a single objection stalls all twelve. Under item-level approval, eleven items move to ordering while you re-source the twelfth.

What every item needs before you send it

A client cannot approve what they cannot evaluate. An item that arrives without context comes back as a question, which costs you the same round trip as a decline. Before an item goes out, it should carry:

• A photograph large enough to judge on a phone, which is where it will actually be opened.
• The price the client pays, in the form they agreed to — not your cost, and not a figure that will change at invoice.
• The room or area it belongs to, so a client reviewing a kitchen is not reading about a guest bath.
• Lead time, or at least whether it is stocked. Approvals stall on the fear of an unknown wait far more often than on taste.
• A one-line reason it was chosen. This is the part designers skip and the part that converts hesitation into a yes.

Three answers, not two

Approve and decline is a false binary, and clients feel it. The most common real reaction to a selection is neither: it is "yes, but can we see it in brass?" Given only two buttons, that client either declines something they actually like — sending you off to re-source a piece that was already right — or says nothing and stalls.

A third option fixes it. Call it Ask, or Needs Review; what matters is that the middle answer lives in the system instead of in your inbox. Programa offers a Needs Review state alongside Approved and Rejected. NYA offers Ask, which lands the client's question in the studio's activity feed still attached to the item it is about.

What to do when something is declined

1. Read the decline as data, not rejection. An item declined in week two is cheaper than the same item declined on install day.
2. Find out which attribute failed — price, scale, finish, timing. A decline without a reason costs you a second guess.
3. Re-source against that attribute only. Replacing a declined chair with something entirely different usually means starting the conversation over.
4. Send the replacement as a new item rather than editing the old one, so the history of what was rejected and why stays intact.
5. Leave the declined item in the record. It is evidence of the work you did, and it prevents re-proposing something already turned down.

Keep the record, because one day you will need it

The purpose of an approval system is not the click. It is being able to answer, eight months later, the sentence "we never approved that" — with a date, an item, and the name of the person who approved it. An email thread cannot do this reliably. A per-item, timestamped, attributed record can.

Send it somewhere they can open

None of this survives contact with a login screen. If approving requires the client to create an account, verify an email and set a password, the process reverts to email on the first evening they are tired. Portals that open from a link — no account, no password — exist precisely because this step is where approval processes die. That question is worth settling before you choose a tool.

How much to send at once

Approval requests fail at both extremes. Send the entire house and the client freezes, because you have asked for forty decisions in one evening and each one feels irreversible. Send one item at a time and you generate forty notifications, which trains them to ignore you.

Send by room, or by whatever unit the client already thinks in. A room is a coherent decision — the client can see the pieces against each other, judge the whole, and finish. It is also the natural unit for ordering, so an approved room converts directly into a purchase order rather than into a partial one that waits.

Chasing without nagging

Silence is rarely a hidden objection. It is usually that your client opened the link on a phone in a waiting room, could not give it real attention, and never got back to it. That calls for a different follow-up than a client who genuinely dislikes what you sent.

1. Check whether the link was opened before you write anything. Not opened is a reminder; opened three times with nothing approved is a hesitation, and they need opposite messages.
2. For a reminder, restate the deadline and the consequence rather than the request — the hold on pricing or stock is the real reason to act.
3. For hesitation, stop writing and call. Something specific is blocking it, and they have not found a comfortable way to tell you in writing.
4. If two items are stuck and eight are fine, approve and order the eight. Letting a whole room wait on one pendant is a self-inflicted delay.
5. Record what unblocked it. The same client will hesitate the same way in the next room.