Behavioral Funnels — Behavioral Design Models
Models to find the shortest path from user to value
This article is the third part of a series that will provide you with a set of models (simple systems to follow) that will help you get from an idea or concept to an MVP definition. You can find the [first] article here and the [second one here]
In my work as a Product Designer, I’ve seen that most products fail because:
- They miss to identify what behaviors they need their users to do to perceive the product as valuable
- They don’t have a strategy to get them to that point effectively.
- And they forget to design to provide value to each one of the roles involved in the product.
The objective of these models is to ask you the right questions for you to design centered on what matters most: driving user behavior to generate value for them.
This third article deals with Models #6, #7, and #8 that helps you find
- The required behaviors for each target behavior.
- An optimized path to reach that target behavior
- A prioritized, and categorized list, to tackle each one of these behaviors.
You can find a detailed explanation, examples of usage, and the templates of the models at behavioraldesignmodels.com.
On the [first article], we have identified the growth, nutrition, and reproduction behaviors required for your product and ensured that they make In the [second article] we tackle the issue of prioritization. This time we will center on how to get one of the many target behaviors we have already recognized to happen by setting up the conditions for that behavior to be possible.
#6 Target Behavior Funnel — Behavioral Design Models
Many steps often precede target behaviors. This model will help you map the required behaviors to get an actor in a position to receive value.
Target behaviors are necessary to make our product show value to an actor. But, we have yet to discover what is required for this behavior to happen. You see, every behavior that we try to create in our actors does not exist in a vacuum. In general, they require that other behaviors have happened before. These are required behaviors.
Imagine you want to ride the metro somewhere; the objective behavior is clear: you want to travel somewhere. There is one obvious requirement the company requires you to pay for a ticket and travel around the city. However, there are several previous behaviors such as:
- Find and locate the station.
- Identify which line and direction to travel
- Wait for the subway at the station
- Notice when it arrives
- Get on it
- Get out at the right stop
If one of these behaviors, for whatever reason, does not happen, the target behavior has no chance of being executed. For example, if you cannot find the subway station, you will not reach their destination. Likewise, if you can’t climb the vehicle’s steps or if you can’t figure out which line takes you to the place you want to go, you won’t reach your destination and probably won’t use the subway next time. Therefore, we need to design all the steps that lead each actor to the target behavior.
Therefore, they should be critical for us as product designers. Without these previous behaviors, the objective behaviors have no possibility of existing.
Some of you might have already discovered that this structure is much like a conversion funnel. A conversion funnel is a term common in sales, a structure where each step of the funnel represents a stage where users might drop off. A conversion (or a deal) happens when a customer travels through all the stages without that happening. In our case, each consecutive required behavior is a step on the funnel.
The model
This model is about recognizing required behaviors. First, write down the target behavior you are currently working on on the right of a page, and circle it down. Then ask yourself: What does this actor need to do this?
Consider your answer; you should find behaviors. For example, “Having Money” is not a behavior, “To Pay” is. It is easy to get tangled and write down resources or motivations as steps in this model. To double-check that you are enumerating an action, look if you are using a verb.
It is also important to note that the granularity we decide on will affect our model. For example, “To pay” can be a single node, or you can break it down into, take the wallet out of the pocket, swipe the card through a device, etc. The first case is correct for most scenarios, but this granularity can be helpful if the product is, for example, a payment device.
There is often more than one behavior that can set up the conditions for others to happen. For example, you can place your order online or at the counter. List all the behaviors you find in a column to the left of the target behavior and draw an arrow pointing at it. Then, consider these required behaviors as targets and ask the same question for them:
You should continue adding columns to the left you reach a behavior
you can consider trivial to the problem we are solving. Of course, defining a trivial behavior will depend on the depth you want to explore the problem, the nature of the actor, and the behavior you want to create.
By the end of this process, you should have a tree-like structure that spawns from right to left of your page and lists all the ways a user may reach the state to perform the desired target behavior.
Example:
My business is selling hamburgers at a fast-food joint. My burgers are the best in the world; therefore, I know they will see value in their purchase if people try them. From the user’s perspective, this is the objective behavior, not to buy the hamburger (which generates value for the business).
What do they need to be able to eat the hamburger? To have the hamburger in the first place, I mean to get it at the counter when the order is ready. And for this? To pay for it. Before that? To choose what hamburger to get, either at the counter or on one of the touch screens. Finally, they must have entered the premises to pick them and drive from their home to reach the door of the premises (let’s assume no delivery option is available). For this scheme, the model looks more or less like this:
#7 Behavioral Funnel optimization — Behavioral Design Models
Optimizing the steps required to get value out of your product is one of the most cost-effective practices. This model will help you identify the shortest path to value and discover shortcuts to save development time.
The previous model (Model #6 -Target Behavior Funnel) explores all the possible ways a user can reach the position to consider doing a target behavior. Since our goal is to build an MVP, it is helpful to consider the minimum viable version of this. This model will help you find a minimum set of required behaviors to design for.
Consider the length of the funnel at equal drop-off rates, the more steps in the funnel, the more users you lose on the way to the target behavior. Since we can only estimate the drop-off rates, we should favor shorter funnels paths over long ones.
Second, the funnel can only have as many users as they enter in the first step. After the first step, users only drop-off of the chain, so we consider which initial point is more convenient.
Finally, it is quite common that, having discovered those behaviors that allow our product to show value, we want to focus all our efforts on polishing the mechanics that support it, since, after all, those are essential for success. But, this is a mistake. The most important behavior to drive on the actor is not the target behavior is at the top of the funnel. If we can optimize this behavior, more people will be closer to performing the target behavior.
The model created before is a directed rooted tree. It means that it has two special kinds of nodes, a single node where no link departs from (there is no arrow coming out of this node), called root. And, several vertices with no edge coming into them, called leaves. The root is the target behavior, and the leaves are our top-of-the-funnel behaviors. Each path between a leaf and the root is a possible set of behaviors to explore. The objective is to find a minimum path that is viable for users to perform this behavior.
The model
Instead of finding the minimum path available from the get-go, we need to optimize the model a little bit further. We will achieve this in two ways;
The first is to go through the tree step by step, asking ourselves, “is there a way to avoid these steps?” Then, if that is the case, we will replace it by finding entire branches (series of nodes) of behaviors with simpler ones.
The second step is to find situations where all the previous behaviors have been reached, but casually, without the particular intention of carrying out the objective behavior; this is what we will call a shortcut. The clearest example of this is when an actor has to drive to the premises. Sometimes he will pass through the premises because he has to do so. Therefore, this preceding behavior is fulfilled with no design effort. If you find any, you should consider shortcuts as leaves!
Finally, measure the steps between a leaf and the root, and select the shortest one. Consider if the amount of people at the top of the funnel is large enough to make the product viable for this first iteration. If not, discard this path and select the shortest remaining one. Next, list all the behaviors from leaf to root. The final result is a prioritized list of behaviors you need for this user to see the value. Now off to design features that will support those!
Example:
Let’s expand the example we used before, “You are selling the best burgers in the world,” and optimize its funnel.
And define some ground rules for this particular optimization:
- Due to the operational restrictions of my fast food place, we cannot do delivery.
- There are no waiters.
- We require people to pay.
As my burgers are the best in the world, I could take all other menu items and eliminate the choice of the order. I can also postpone the payment until after the target behavior and relate it to a mandatory behavior, such as leaving the premises.
These might not necessarily be the best options for operation, but it makes it easier to try the burgers and validate that we have the best burgers in the world. Building this diagram helps us question how to revolutionize my business and bring my potential guests closer to my product.
Let’s try to find shortcuts; I can take advantage of customers already near the premises and add posters that invite them to enter, no need to bring people from their homes. I can also take advantage of those customers who have already ordered and include a hamburger with any order they make.
So the new diagram would look like this:
#8 Features that Drive behaviors — Behavioral Design Models
How to design features to drive behavior? This model will introduce you to a new way of thinking of behavior design and classify your target behaviors in terms of their weakest elements.
elements.
We are finally ready to focus on a particular behavior and how to get people to do it. This model is the model you need to create for each behavior in the path of your target behaviors. So I’ve kept it simple.
Whenever you want to get somebody to do something, there are three scenarios:
- They don’t want to do it (lack of motivation)
- They can’t do it (ability/friction)
- They don’t realize they can do it (triggering issues)
These three dimensions shouldn’t be considered as absolutes but as gradients. For example, I could not be motivated enough to go to the gym, but a gentle push can get me to go, or maybe there is no way in hell I’m returning to the gym, no matter how much money you are offering for me to do so. Each particular user will have their threshold to achieve for the behavior to happen.
Pre-product, it isn’t easy to estimate what those thresholds are. After you collect some testing data, you will be able to gauge these realities better, but for now, we will have to use heuristics to focus our design.
The three dimensions I introduced earlier (motivation, ability, and triggering) are part of B.J Fogg’s behavior model, a simple and effective way of looking at behaviors.
Fogg’s model states that: “ three elements must converge at the same moment for a behavior to occur: Motivation, Ability, and a Prompt. When a behavior does not occur, at least one of those three elements is missing.” it also states that there is a threshold that the conjunction of motivation and ability must reach at the time of the prompt for the behavior to happen. And that threshold has this shape:
So if at the time I see an ice cream truck (triggering) I’m having a craving for sweet(motivation) and got some extra money, the time to queue, and the truck is not that far away (ability), I will buy myself ice cream!
This model is a great way to explain what factors are in play when designing for behavior, and I strongly recommend you visit his page https://behaviormodel.org/ to know more. But how do we use it in our quest for an MVP? by now, you should guess my response, prioritize!
The most significant dimension is triggering. Without the user noticing the prompt, he doesn’t even consider the behavior. In “models for triggering design,” I will give you some tools to design triggers and how to get users to consider a behavior (even unconsciously).
Then we have to decide if motivation or ability. Counter to what your gut might tell you, we should consider ability first. The reason is relatively straightforward: it’s easier for two different people to agree that something is difficult or expensive than to agree that it’s fantastic or engaging. So designing to reduce friction (or boost ability) helps you target more people and get more of them closer to the activation threshold. Check out the models for friction reduction and learn some more.
Finally, let’s talk motivation. Motivation is hard. Motivation is sexy, and everybody wants to know how to get people motivated. I have been talking about motivation since model Model #1 — Product and Life. In “Behavioral Design Motivation Models,” I will help you navigate these waters, but you must first focus on the other two dimensions. There is no yearning for something you don’t know exists, nor is there a way to take a weekend vacation to Mars, no matter how much you want it.
The model
Pick three colors, one for each dimension. I use green (triggering), red (ability), and blue (motivation). Go over the list of behaviors and think which the most pressing questions are:
- How/when would they notice? Would they understand what to do? (green)
- Is it simple? can you ensure they have the resources needed? (red)
- Will they want to do it? (blue)
Mark each behavior with the color, and if you marked it blue, I would recommend you double-check! Then open the corresponding model and brainstorm some solutions to that key question.
These three models will output a complete list of behaviors you need to design for on your MVP.
For the other articles of this series, go to:
The future
There are still at least three models to come. These will help you work on each one of the dimensions and help focus the feature on boosting that weak aspect.
Subscribe to my medium (https://nachoparietti.medium.com/subscribe) or follow me on my social networks to be notified when these articles pop up!
Here are a set of quick steps to construct the models.
Model #6 — Target Behaviors Definition
- Write the target behavior to the right of the paper
- Ask yourself: What does this actor need to do this?
- Add the behaviors to the left and link them to the existing behavior
- Check you are adding behaviors, not requirements
- Repeat from 2 for each new item you add until you find trivial behaviors
Model #7— Behavioral Funnel optimization
- Simplify the model. Go over each node and ask: is there a way to avoid these steps?
- Find shortcuts. Situations where all the previous behaviors have been reached without the particular intention of carrying out the objective behavior
- Find the shortest path from a leaf to root, the minimum path for that target behavior.
Model #8 — Behavioral Funnel optimization
- Go over each behavior and consider the most likely scenario:
- Mark it Green if they don’t realize they can do it (triggering issues)
- Mark it Red if they can’t do it or it’s hard to do (ability/friction)
- Mark it blue if they don’t want to do it (lack of motivation)
Hi, I’m Nacho Parietti
I help design products that drive behavior at ingenious. Over the last 8 years, I have designed and optimized products, on adoption and retention, under a behavioral sciences lens.
I’ve learned and evolved some of these applied techniques during my time working for gambling and gaming industries, some from while tens of products for different types of startups, and others while translating research into real products for researchers in leading universities.
If you are interested to learn more, go to behavioraldesingmodels.com or feel free to reach out! LinkedIn / Twitter
