Amazon is the best example where owning the full delivery stack is important. They can offer the widest selection, lowest price and fastest shipping but if the experience of finally receiving your order is poor, that will reflect badly on Amazon. If your order was late, lost or damaged, it doesn’t matter where the true fault lies – Amazon are to blame. That’s why they’re now running their own logistics end to end.
This principle applies widely and yet few seem to understand how important the end to end customer experience is.
Meetings are a good example where there is an incentive to automate out the hassle. Going back and forth to find a day and time that works for two people is difficult. Timezones. Locations. Changing schedules. And it seems to get exponentially difficult the more participants there are.
Proposed solutions range from AI assistants through to selecting time slots from a web calendar. A lot of “busy” people use human assistants as well. But I’ve always found these unsatisfactory. AI assistants have made “mistakes” which are embarrassing to correct. Sending people to a link seems impersonal and handing it off to an assistant is only as good as that assistant – if they’re slow to reply or make errors, the whole process becomes frustrating. For some this might not matter much, but if meetings are a core part of how you do business e.g. investors, a poor experience might mean you lose deals.
The same principle applies to software. Onboarding customers to SaaS is a critical step in ensuring new customers successfully deploy, which is key to preventing churn. Onboarding is well understood but there is much more to the software experience, particularly for developers. Documentation, API references and programming libraries are all part of the “first run” experience. MongoDB is one of the most successful examples of this.
Mapping out the full delivery stack and having someone own the full end to end experience is the only way to keep quality high. Over time, the ownership and responsibility of each component will be delegated to different people and teams. Decisions will be made that make sense in isolation but affect the overall flow, and they won’t be spotted unless someone regularly runs through it all. Even the CEO should regularly test their product, as Bill Gates found in 2003.