This is a common advice given when you start a business. In order to get happy clients, you need to under-promise and over-deliver. But when trying to grow a serious technical consultancy, I think it can be one of the most toxic advice you will receive.

If your aim is to get some business as a one off with a client, and you know you will never see her again, then maybe, you can use the under-promise / over-deliver proposition as a marketing stunt, banking on the fact that a happy client might recommend you to someone else.

Now, if you’re growing your business, you’re not really looking for one-off clients. You should be looking to develop a healthy and steady relationship with your customers.

And this is exactly when the under-promise / over-deliver proposition becomes problematic.

I want my free sweeties

Imagine that, on your way to work, you discover that a new bakery has opened. You enter the shop and order one large cookie. To your surprise, in the bag with your cookie, the baker also throws in a handful of sweets.

Sweets are cheap to buy for him, they do not require any preparation and it makes the new shoppers happy. They also are loaded with sugar, which has this side effect of subtle addiction.
More importantly, the baker’s cookies may not be that different to cookies in any other bakery, so it might allow him to add a differentiation factor.

Obviously, you’re happy about this surprise, and you will come back. You might even tell your friends about this fantastic baker throwing a handful of sweets when you buy a cookie.

On your way to work, you now stop every day at the bakery, for your morning cookie and some added sugar. You also realise that the word gets around, and you are not the only one waiting for your daily deal.

After a few months, the success of this operation is so big, that the baker is facing an issue. The costs of sweets are lowering his margins, and he is now facing 3 options:

  1. Reduce the number of sweets he gives away, from a handful to a couple. It will still impact his margin, but in a less significant manner.
  2. Keep what looks like a good deal cookies + handful of sweets, but raise his price. His margin will improve, but some customers may be put off by the price increase and stop buying cookies from him.
  3. Completely lose the sweets and keep the cookie at the same price. Again, his margin will be better, but some customers might get frustrated to pay the same price for a reduced product.

Free features are like sweets

When under-promising and over-delivering is the de facto way to operate for a company, it quickly becomes toxic because:

  1. In the same way that sweets are addictive, delivering more features than planned are also addictive for the client. If one day you can’t over-deliver, it can lead to all sorts of questions.
  2. In the long run, it risks having impact on your margins. If you always have to push more features than expected, you might put yourself in a tricky situation where you will have to put more people than expected on features.
  3. Bad mouthing goes around even faster than praise. If you nurture your customers with over-delivering, when the time comes when you can’t do it anymore, you risk facing the word or shame: “Oh, they used to be good, but now, it’s not the same anymore”.

How can we solve this

If we go back to the example of our baker for a minute, I don’t think that the main issue was with the free sweets from the beginning.
If the baker’s cookies were the best in town, word-of-mouth might have worked the same way. He might even have been able to raise his prices from the beginning, knowing that he had the best cookie recipe, and maybe throw a couple of free sweets to his best customers from time to time.

In our industry, it all comes down to our work quality, experience and estimation. Estimating features is an art, and it requires experience and good context knowledge.

The message we are sending to our customers is that, most of the time, we know what we are doing. When we estimate a feature or a set of features to take X days, it’s most likely to be accurate.

If the feature set puts us in the unknown, then we need to be open and clear about this with the client, and we usually put risk factor attached to our estimations.

Be accurate

Estimations are hard, and accurate ones even harder. They require a profound analysis of how you or your company works.
You will need to deeply look at each person in your company dynamics, what type of workload put people in ‘stress mode’, what type of quality you want to ship, and finally, you will need to trust people’s judgements.

Accurate estimations should be part of any developer training programme. Start small, with features or clients where you know that an incorrect estimation will not be easy to explain. Then, gradually expand the range and complexity of features to estimate. Experience is the only way to improve this skill. And accuracy and openness are the keys to a healthy relationship with your clients.