For years, you've heard that training yourself or others to a desired habit requires some positive reinforcement--that is, you need a reward to look forward to so your brain squirts out dopamine, you feel good and motivation stays high. And on a basic level, this is 100 percent true.

But maybe you've noticed that Joe Schmoe's performance is waning despite the fact you're telling him every day his work is fabulous.

Or maybe, after being on a diet for a while, that treat you give yourself every single Saturday at 2:45 p.m. has lost its deliciousness.

But....but...rewards are supposed to work...

The issue likely isn't with the reward you're giving, but when you're giving it.

Your brain on randomness

When you give a reward in a very predictable way, your brain, which adores novelty, quickly learns what's coming. It knows exactly what's going to happen, and so the reward centers in your brain essentially say, "Meh, this isn't stimulating enough anymore, I'm bored, and it's not worth pushing the dopamine."

But when you never really know when a reward will land on your desk, the brain can't figure out whether there's a pattern going on. There's no rhyme or reason to getting goodies, so your brain stays curious and is willing to keep the dopamine flowing to see if the expected result happens.

Now, the caveat here is that you cannot start with a totally random reward system. First, you have to teach your brain to make the association between a specific behavior and a given reward. For example, say you want employees to be better about finishing work on time. You could give an employee an extra $50 for finishing a project early, and then make sure to give that same reward for the next half a dozen projects or so that are ahead of schedule.

But once the association is clear, it's time to shake it up. Because your employee has learned that work finished early gets a bonus, you can start giving the rewards intermittently. Your employee will keep trying to work fast, because although they don't get a "win" every time, they still get occasional confirmation that a win can happen and that the pre-learned association hasn't broken down. To their brain, it's still worth trying.

If this "variable ratio schedule" system seems a little familiar, it should--casinos use it to keep you coming back to the slots over and over. And psychologist B.F. Skinner, who researched different reward schedules, found that this approach gets people doing the desired behavior more often than any other. What's more, he also found that it yields behaviors that are stickier--that is, that are harder to stop--even when the rewards aren't coming anymore.

The dark side of no predictability

Importantly, while there are dozens of ways to use intermittent rewards positively, it's not uncommon for less scrupulous people to use them manipulatively. For example, a toxic boss might treat you fabulously when you first join their business, teaching you to expect kindness. Over time, though, they might start treating you worse and worse, always showing on praise and opportunity every time it looks like you might actually leave.

And from another perspective, intermittent reward connects to your view of risk. For example, you might cognitively understand that the stock market isn't a sure bet. But if you've constantly received good returns, you likely will be more willing to put money into shares and ignore market warnings, because the reward of profit has happened to you so often.

The bottom line is, consistent, predictable rewards are best for establishing new habits. Intermittent rewards are best for ensuring you continue those habits over time. Always be aware of and open about what the goals are and make sure that you and others are using intermittent reward for the right purpose.