Automation’s Impact on Product Management

First impressions.

Automation of different aspects of our current lives has changed (and will continue to change) many workplace roles and how things are done. The role of product management will also undergo a major transition due to automation.

Customer expectations of new products are changing drastically which in turn changes how products are conceived. For instance, as a driver, I care about in-car controls such as the steering wheel, lights, indicators, and wiper controls. This changes once I am sitting in the back of a ride-share or self-driving car; I do not care about driver controls but only about reaching my destination safely and on time.

The controls are necessary to help the customer achieve the end goal but they are not the product. There is a saying that when customers drill a hole, they are looking for a “half-inch hole,” not a “half-inch drill.” Controls without the necessary automation are a necessary evil, but do not serve a purpose toward the end goal or use case of the customer.

As we move toward automation, there are a few things that will have to change regarding how products are designed:

Decision Making Shifts to the Left

In product management, one of the key questions that need to be tackled daily is how a specific feature should be delivered to the customer. The usual answer is that we are not sure how the customer will use the product, or there might be a corner case where this feature might have to be used differently. Thus, the easiest solution is to create a “knob” and let the customer decide.

This phenomenon has resulted in a large number of product features that are available when we need them. The reality is that the majority of customers never use more than 5–10% of a product’s capabilities. This also moves decision making from the product manager to the user.

As the world moves toward more automation, decision making will have to occur earlier and cannot be left to the user. In an automated system, machines will make decisions in real time based on the available information. The responsibility of product managers and engineers will change: they will have to consider every possible scenario, and use machine learning (ML) and data to train machines to react to various situations. It is no longer sufficient to simply add a “knob” to a product; one will also have to consider the knob’s setting based on current conditions.

Value Proposition Shifts to the Right

Today when someone looks to buy a car for their own use, they care about brand, driver comfort, and capabilities available in the models. The opposite is true when it comes to ride-sharing: I do not remember even half the models in which I have traveled, as the goal was simply to reach my destination. As things become more automated, it will become more difficult to differentiate products based on controls/knobs, which had been the main area of focus until now.

If an automated product works and does its job consistently, users will be inclined to use it. In order to stand out, the value proposition will have to transition to higher-level functionality and focus on new areas that are important to the customer.

"In order to stand out, the value proposition will have to transition to higher-level functionality and focus on new areas that are important to the customer."

Sanjay Kalra

New Feature Requests Will Decline:

In my previous company, we focused on automation and removed a lot of manual knobs/rules, thinking this would help customers. The biggest surprise was that removing manual rules also helped us. If the customer has fewer knobs to tweak, consequently they will need fewer features. In the same example of the self-driving car, if I am not driving and using all the driver controls, I will not ask for enhancements to those controls. Even so, as products will continue to evolve in different areas, I am sure users will certainly think of new features to help them keep busy or make life simpler.

Escape Hatch:

The Boeing 737 is a perfect example of what can go wrong with automation. When the planes had problems, humans were unable to take control, with tragic results. It will be a long time before machines become perfect and reach the level where they are trustworthy in all scenarios. Any automated product being designed will have to take into account what happens in case of failure. How will humans be able to take control and accomplish some of the critical tasks that machines are supposed to execute?

Value Realization Will Have to Become Explicit:

One of the challenges in an automated world is that if things work as planned, you get used to them and stop noticing them. As there is no human interaction, the only time the customer realizes the value of the product is when the product stops working or has issues.

In an automated product, it will be critical to make sure the customer realizes that they are receiving value from the product. The pre-emptive messaging and regular interactions with the customer will become even more important. Customer Success will play a critical role in interacting with the customer on a regular basis to ensure “Value Realization.”

We are far from science-fiction machines designing other machines/products. Product managers in the new automated world will be responsible for making more decisions, not fewer. They will have to get used to working with ML/AI and data to help them do their job as decisions can no longer be left to the customer or made instinctively without supporting data.

Pain Killer - What type?

Pain Killer – What type?

When I was starting my company one advice, I received repeatedly was that build a pain killer and not a vitamin. The logic is that customers are willing to buy a pain killer at any price and from anywhere as long the new product is proven to be the solution to a pain point. In comparison a vitamin is like discretionary spending and people tend to procrastinate or all together skirt non-essential spending.

I agree with this general theory, but I didn’t get any insight about what kind of pain killer to build. There are two kinds of pain killers one for every minor aches and pain e.g. Tylenol which is available without any special prescription and is a part of everyone’s first aid kit. These kind of pain killers are not differentiated from one another and requires significant amount of marketing but have a simple use case which every customer understands.

The other pain killer for severe pain e.g. Morphine which is available only with prescription and it is required only under special circumstances. The product in this category are typically very differentiated but not needed every day, so customers need to be sold on the value of these especially when they have not experienced the extreme pain.

In security market too the same logic applies as zero day and sophisticated attacks which are like severe pain points which do not happen every day but common compliance or protection against known threats is needed by every customer.

The ideal solution in my perspective is to build a platform to be successful in long term but market specific common pain features like Tylenol to get early and large-scale adoption but have Morphine (as part of the platform) to differentiate.