When a team says the onboarding needs work, most people picture three illustrated screens that appear after registration. Those screens are the smallest and least important part of it.
Onboarding is everything between seeing the product for the first time and getting something useful out of it. By the time someone reaches a welcome carousel, most of the decisions that determine whether they stay have already been made.
The chain is longer than the product
The path a person actually walks starts outside anything the product team controls, and every link in it can drop them. Writing it out is uncomfortable, which is exactly why it is worth doing.
Each step is a place where someone can fail to understand where they are, why it matters, or what to do next. Teams tend to instrument the last two steps and guess at the first six.
- Sees the ad, post or recommendation
- Follows the link and arrives somewhere
- Works out what the product is and whether it is for them
- Enters the product, or installs it
- Creates an account, or is asked to
- Reaches the main screen
- Tries the first meaningful action
- Gets the first result worth having
The problem is usually not the traffic
When the numbers are weak, the default reading is that there are not enough people at the top. More ads, more subscribers, more reach. Sometimes that is true. Often the volume arriving is fine and the product loses them before the value lands.
This misreading is easy to make because acquisition is measurable and the middle of the path frequently is not. A funnel that reports installs and purchases but nothing in between will always look like a traffic problem, because the only two numbers you can see are at the ends.
You do not have an acquisition problem until you can prove the people arriving are reaching value.
Why this work pays three times
Fixing the path to first value is one of the few changes that improves several numbers at once, because all of them depend on the same event: the person understood what they got.
Conversion improves because more people reach the action that counts. Retention improves because someone who understood the product has a reason to come back. Monetisation improves because they were still there when the offer became relevant.
None of that requires buying anything. The people are already arriving.
What a good path looks like
The rules below are unglamorous and they are where most of the recovered value comes from. They are also the ones most often traded away, usually to satisfy a data request from another team.
Notice that four of the five are about removing something rather than adding it.
- The person understands where they are within seconds, without decoding a slogan
- Value arrives before the product asks for anything in return
- Every step makes the next step obvious
- Nothing is requested that the product does not need yet
- The sequence ends at a result, not at a created account
It is usually a sequence problem, not a design problem
In most of the products I have looked at, the drop-off was not caused by how a screen looked. It was caused by order. A question asked before the product had earned the right to ask it. A permission request in front of the thing that would have justified it. One extra confirmation step that existed for internal reasons.
Each of those reads as reasonable in isolation. Stacked in sequence, they ask a stranger to keep paying small costs before receiving anything, and a predictable share of them stop paying.
This is why I map the sequence before touching a layout. Reordering is cheap. Redesigning the wrong step is not.
Onboarding does not end when someone registers. It ends when they can say what the product is for and why it is worth returning to. Until that moment, every earlier step is still on trial.