You can spot a breakout hit in four hours

When launching something new, you form a hypothesis about how it’s going to perform. You pick a feature that you think will move the dial, or you write the announcement you think people will read. Then you wait.

I’ve been writing Product Notes for over a year, and I still don’t always know if a post will land. Sometimes, I’ve written what I thought was a great post, and it doesn’t resonate. There have been times when I’ve thought twice about sending a post because I wasn’t sure, and some of those have been really popular.

What has changed is how quickly I can tell if a post I’ve sent is doing better than average.

It’s hard to predict what will resonate. You’re guessing at what will grab someone’s attention. Finding out if your prediction was correct is slow. If you’re launching a feature every month, you won’t build up enough experience to trust your own judgement on it.

Detection is different. You aren’t trying to answer whether a launch will be successful in the long run. You’re looking for early signs of whether the launch is working. It’s a narrower window, so you find out faster.

The problem is checking a week later, and finding out that something over or underperformed. The moment when a bit of extra effort might have compounded has passed.

At Product Notes, I check the views for each new post after it goes out. For instance, I might see 26 views in the first few hours. The number doesn’t mean much on its own, but when I know what other posts were on at the same point, I can compare them.

For comparisons to be possible, you need a set of reference data. It’s months of recording numbers for things that may have been average, good, or bad. Keeping a lot of numbers isn’t that exciting on its own, which is why it’s easy not to do it.

Platforms such as Substack make it easy to track the total stats of a post overall, but less easy to answer “where were all of my posts sitting, four hours after publication?”

I now have that data, and can use it to spot if a new post is accelerating faster than usual.

My last post was on internal admin screens. It had 26 views and a 22% open rate at the four-hour mark. This was after sending it to my subscribers, but before doing any other promotion.

The previous post took a full week to reach 34 views. So four hours in, this one had already pulled roughly three quarters of that.

I noted the pre-promotion figure first, then cross-posted to LinkedIn. The order matters - promote first and you’ll never separate the LinkedIn traffic from the traffic the post was getting anyway.

The admin screen post ended the week on 49 views. LinkedIn was responsible for some of the bump. So it was worth promoting.

It’s easy to do this backwards. Your subscribers get every post anyway, so promotion isn’t really for them. It’s for people who haven’t chosen to hear from you - a much smaller pot of attention. Push every post equally hard and you’ve told people nothing about which ones are worth their time. The one that genuinely deserved it gets the same reaction as the rest.

Publishing should be cheap and consistent. Promotion shouldn’t be either. Save it for the things that have shown you something, and then go hard while the window is still open.

Promoting on publication day is promoting blind. Promoting something that has already demonstrated it can hold attention is a much better bet, and it costs you a few hours of patience to find out which one you’re doing.

While it does show where to spend promotional effort, it doesn’t tell you if the post will actually land with a wider audience.

From my older stats, I found a post from July that I’d logged at 39 views and a 24% open rate, which felt like a miss. However, the post still got readers some time after its publication date, settling at 59 views and a 29% open rate. Other posts have done the same thing.

The verdict I’d come to in week 1 was premature. Many posts keep growing, long after they first went out.

There are limits to how I’ve been tracking my posts. I’m comparing a day 1 figure against the 7-day totals of previous posts. This isn’t comparing apples to apples. It does help to spot a clear outlier, but it doesn’t do much more yet.

The way to get better is to build that reference data. I’ve always logged total views and open rates, updating them if they change, but I wasn’t tracking how posts looked on day 1. I’m working on that now, and it’ll take time to be useful.

The advice here isn’t about getting better at predicting. If you want to spot the breakout hit early on, you need to start recording similar “early stage” stats for your launches. Once you have that data, you’ll be able to see much more quickly if your launch is an early success - and whether wider promotion is worth a try.

Product Notes is free and lands every week. Subscribe to get it in your inbox.