Ask HN: How do you convince a client who doesn’t want to change their mind?

[alexsexotic]

Hi

Iam sure that this is a common problem, but for those who have experience with dealing with this.We are working on a project that is taking almost over 2 years, with few user testing and real world production exposure. One of the reason is that our clients demands alot of absurt features with no or few validation backed up and we try to talk him out of it and reduce the number of features to the min. needed.

Recently we had this problem with the landing page which was in its 2 remake and we communicated with him about the changes and listened to his feedback. He agreed with the layout, colors and text content so we finallized that spend an amountible time on it. In the end he send us an email about his imagination of how it could look like AND it was totally of from what we agreed on. We even got some feeback from users about the landing page, which was fine.

So my question is how would you deal with it? What could we done better? For example find a person who can talk over better the client ?

[gt565k]

This is really partially process and payment related question.

This is why you make design and deliverable requirements and have the client sign off on them during the RFP phase.

Payment terms structure and estimate padding for the project cost should account for risk of client changing requirements and/or completely backing off.

Example payment structure:

15% due at start of project 20% split in between in 4 payments as time goes by 5% due at final deliverable

Payment terms are generally flexible and vary depending on your ability to estimate work and risk. As well as what the client is comfortable with.

Once you make the client aware that they have to sign off on requirements and make payments at specific time intervals during the project, they will be more involved in making sure what they want is conveyed and there's no miscommunication.

[brudgers]

It is also an expectation problem. Not the client, but the contractor/consultant. Design is iterative. Engineering design is like engineering: it's reasonably objective. Most other design is not because better color/animation/layout/texture/etc. is not reasonably objective because the criterion is "what is the client happy paying for."

The tenor of the question suggests the degree to which expectations are off base. You can't convince a client that they like a different color scheme; three columns over a sidebar; tabs more than endless scroll. That's just how it goes.

I completely agree that it is process and payment related, too. The process and payments are a result of expectations.

[codegeek]

After a certain point, you stop trying to convince them. You start charging them for every change regardless of how silly the change is. Make it clear to them that every change is chargeable/billable and they will stop.

"He agreed with the layout, colors and text content"

How did he agree ? Was it in writing ? Were exact designs done or was it more of a verbal discussion ? With clients like these, you need to be as detailed as possible.

The moment they add a new change, you say "This is a change request and here is the estimate for it: $x and timeline n additional days. Once you confirm, we will proceed"

If you start charging them, they will either keep giving you money or they will stop asking to change things.

[itake]

As a freelancer myself, you're paid to build the product and make sure the client is happy. It is their job to design the product and to find product market fit.

I have watched so many people shoot themselves in the face over things that I thought were dumb, but did it for them anyways because they asked.

Yes, I know it sucks to have a portfolio of bad clients. Really the only thing you can do is fire the client and find someone that is easier to work with that has better potential for a successful project.

2 years seems like a long time to work with someone. I'd just look for a new client.

[blaser-waffle]

"Do unto others what gets you paid."

Come up with something concrete -- numbers / stats, or documented vulnerabilities, or well known failures, etc. -- and costs related to those. No one cares about issue [X] until you explain it to them that it'll cost [Y] to fix it later.

Then document their answers, make it clear you advised them for/against something, and then do whatever they're paying you for. If they tell you to put a toilet in the kitchen, and they're adamant about it, they can have their kitchen-toilet.

[aww_dang]

One layout for landing pages doesn't sound like the right solution for something which has 2 years of dev time invested. Split testing and multivariate testing is essential for landing pages.

Why argue about the merits of a lander when you can use metrics?

[meiraleal]

You make it more expensive for him. If that is still what he wants, both are happy now.