Stuart Silverstein Experience Strategist/Designer

back to all articles

Optimizing the Funnel

Rule number 1 in web design: Every project has way more ideas than time or resources.

No matter how big,or how much money the project costs, you will always have more ideas than resources. The difference between a successful project and a mediocre one is always in the execution. I’m sure that the vast majority of mediocre websites started out great, and probably included some truly innovative solutions, however, then …. reality sets in. Product team, meet your resources, and that beautiful product you have been planning, gets hacked to a mere shadow of the career topping project you had so much hope for. Bit by bit it gets widdled down to something buildable. And often it’s hacked to the point that the original vision for the project is just some sort of half usable UI, that barely passes for MVP. The wrong features get built, and the project has little traction, and makes little difference in the bottom line. The decision makers loose interest and you move on.This story replays itself over and over for teams without a clear UX strategy.

The above story is a tale in accumulating UX debt that you have to pay with compounded interest. One way I have found to help bridge the gap between a ton of ideas and something buildable that results in the biggest increase in attaining KPI’s is a technique called “Funnel optimization.” Typically it is a marketing visualization, however in my UX approach whereby the strategist outlines the funnel, assesses the funnel, compares the ideas on the backlog with the effectiveness of the funnel, and prioritizes improvements. It helps take some subjectivity out of backlog grooming in order to make improvements in the areas that your application really needs, versus bolstering areas that are already performing well.

The funnel metaphor is always good because it is analagous to real life. Each of the funnels have holes in them that leak out water. There may be holes near the top that are really large, which may cut out half of the water getting to the lower levels. Or, the opposite may be true, where you are getting all the water to the last funnel only to lose at the bottom of the funnel. The same is true for any website – especially a commerce/transactional site. The consumer goes through the different stages and all stages need to be optimized. It just may be that the stage you are throwing your resources at is the one that is already the best optimized. Then you are wasting your resources. The trick then is to find the least optimized points in the funnel and then tighten those up first.

Defining the funnel

Since the 1960’s, a classic method of defining the sales funnel came from a Frank Hutchinson Dukesmith article called “AIDA.”

AIDA stands for

  • Awareness
  • Interest
  • Decision
  • Action

This model describes how consumers move from step to step, advancing through the purchase funnel getting closer to sale, and ultimately retention. It is the best model I have found in helping visualizing steps in an e-commerce process (thanks Tim Ash!) as this model illustrates a method of assessing the interaction design of the site in a very transferable way,. When you add “Satisfaction” as a last step in the funnel, you ultimately close the loop and can accurately assess the funnel from an experience standpoint.

The steps are defined as follows:


Attracting the attention of the customer. This is the goal of most advertising and Search campaigns, however, the info plays a key part in defining the experience. The user comes with knowledge and a problem, and the experience designer must understand both thoroughly in order to design an effective solution.

User Questions:

  • Does a product exist that meets my need or desire?
  • If I know it exists, is my image of it intriguing?

UI Solutions:

  • Landing pages, and the ability of the user to grasp the Unique selling proposition instantaneously.
  • Banner ad integration, and the role it plays in the acquisition experience.
  • The ability of the creative to grab the consumer, both aesthetically and in messaging.



This is the point where the user looks for information on the product. This is the stage where the consumer is doing information gathering. For designers, in order to satisfy this need, we must have a good understanding of what knowledge they have in their head versus what you need to supply them with. The user must know that they have the need and decide to buy.

User Questions:

  • Does this do what I need it to do?
  • What does it cost?

UI Solutions:

  • Clear product pages, knowing the user’s needs, and demonstrate a knowledge of the user
  • Great photography and media access
  • Curated products and top level pages
  • Email capture for more information/offers
  • Cross sells



This is the point where the user has realized they need the product, and is ready to buy, but needs help deciding on “which one?” This may be between competing companies (Dell and Mac), various models – such as a Macbook Air and a Macbook Pro.

User Questions:

  • What options are available?
  • What does it really cost?
  • What are the differences between product x and y? The advantages? The disadvantages?
  • Is there anything else I need with this to use it?

UI Solutions:

  • Comparison features/information/shopping comparison engines
  • Refine/sort/parametric features
  • Mini carts showing Total costs – including extras like shipping and service fees
  • Local/personalized information prior to purchasing, such as deals in your area.
  • Sales Support – live chat, phone/email support, etc
  • Reviews
  • Options
  • A dealer locator to see where I can buy it
  • Upsells



This is the place you really don’t want to mess up. The consumer is now ready to purchase and all you have to do is take the money. This oftentimes is complicated, due to different types of payment, gift cards, discounts, sales etc. All of which need to be effectively communicated to the user prior to completion of the process:

User Questions:

  • Where/How can I purchase this?
  • Am I getting the right price? (discounts, sales etc applied)
  • How do I pay for this?
  • Is there anything fishy about this that may cause me to mistrust this transaction? Will anything bad happen as a result of this transaction?

UI Solutions:

  • Checkout/Registration
  • Invoice/Cart totals
  • Apply discounts/alternate payments
  • Return/Support policies
  • Purchase support – Live chat



This is the last step where you either retain or loose the customer. If the customer was satisfied, the chance of a repeat purchase increases.

User Questions:

  • Did I get what I thought I was buying?
  • Was the quality of the product and product experience up to snuff?
  • Does it actually do what it claims? How well? Was it worth the cost?
  • What was the purchase process like?
  • How do I get support on using this product? Is it adequate?
  • Do I have a need for more products to add with this purchase (additional or refill)?
  • What happens if I have a problem?
    • Was it easy to find more relevant products I aim interested in?

UI Solutions:

  • Confirmation emails
  • Autoresponders – encouraging further action relative to relevant info
  • Personalized suggestions
  • Thank you emails
  • Additional offers
  • Product Support/Quick start
  • Brand engagement apps like games
  • Upsells/cross sells


Translating the funnel to UX strategy

The next step in the process, is to translate all the actions on your site that move a user down the funnel. No doubt there will be primary actions – make a purchase, sign up for email. Find a product, etc – and secondary tasks – find out reviews, ask a question about a product, etc. If you have a rich e-commerce experience, this list can be pretty large —50+ tasks — on a mid range site, you should have about 20-30 that makes sense. Any place you can consolidate and remove functional duplicates do so. For instance, if one of your actions is “Repeat visits demonstrating engagement” and “Monthly repeat visits showing loyalty” – each of those actions might have some crossover in terms of analysis. If you have different user groups or personas, each with unique actions in the funnel, you will need to break each of those tasks per user out.

After assigning each of the tasks, you will want to rate the effectiveness of the site in helping the user achieve his goal on a scale of 1 to 10, 10 being best. (I never give 10’s except in very rare instances). This rating scale must be bought into by all stakeholders or the technique does not work, so it is paramount that the ratings are discussed and agreed on prior to moving forward on the strategy

Here is a guideline for interpreting numbers:

8-10 User understands most or all of the information necessary, and is able to complete the task with no difficulty. A 9 or 10 indicates a level of intuitiveness, grace and sophistication that is exemplary.
6-7 A users can accomplish their task with a little difficulty and may or may not be able to understand all information presented in completing the task. The user will definitely have to think here and might stumble and recover to complete.
4-5 A user completes most or all of his task with difficulty
2-3 A user completes most his task with significant difficulty, at the same time a large constituency would not be able to complete their task.
1 Most or all users would not be able to complete their task at all.
0 Task is unable to be completed because the feature to complete the task does not exist.

Sample action list:

Clickthrough from adult targeted media
Wishlist email click through/opened
Visits to a store/product detail page(2)
Clickthrough to a 3rd party retailer
Share content (2)
Like it/Want it (2)
Regular Monthly visits (2)
Online Purchase
Registered user signed up to receive promotions
Purchase at a brick and mortar store(2)
Repeat purchase

Repeat visits from Wishlist


(2) = 2ndary action

Pulling it together

So by now, you should have your funnel defined, and your actions defined and rated. Now it’s time to put it all together. The next step is to assign each of the actions to a step in the funnel. Go through each of the actions and assign it to a funnel level like below. Calculate the average score in order to see where the biggest holes are and highlight anything that is a primary task with a score lower than 5.


Clickthrough from adult targeted media 3
Visits to a store/product detail page(2) 7
Average Score 5



Wishlist email click through/opened 6
Clickthrough to a 3rd party retailer 6
Registered users signed up to receive promotions 0
Average Score 4.5



Share Content (2) 0
Like It/want it 0
Average Score 0



Online Purchase 7
Purchase at a brick and mortar store using online
info such as wishlist printout or find a store(2)
Average Score 6.5



Increased Regular Monthly visits(2) 4
Repeat purchase 7
Repeat visits from Wishlist 5
Average Score 5.33


As you can see above, this site is having issues getting users to sign up for promotions, like/want a product and click though from adult targeted media. Once you have this info, you can either brainstorm ideas on how you could help the user do those actions, or see any ideas that best attack those problems. For instance, you can add social media like buttons or add a systemwide like/want system where users can see all their likes in one place. Those would be good use of resources since the business desire is to get users to like content or show an indication of intent or preference. Likewise, for the “click through from targeted media,” this means there is a problem that happens when a user arrives on the site without messaging to answer their question. Maybe a sections for adults or a landing page dedicated to handle this media might best help this portion of the funnel.

Conversely, the 2 levels of the funnel with the biggest holes are level 2 and 3. This site is doing a good job of helping users purchase, but most don’t get that far. So another approach would be to see how best to improve the average scores of the funnel levels. This can be done by brainstorming functionality that would address that issue or by pulling from the backlog of ideas that need to be worked on.

Finishing the job

So you have your priorities and your functions so you can start your next sprint planning with a focus. You still need to get technical timing, but once you have an LOE attached to your prioritized list, you can start to make some decisions and plan. You want to focus on features that improve the funnel and measure against it. It is important to make sure all your actions are measurable and co-ordinate effort with the Business intelligence team/analysts in order to see that measurable improvements are being made. The analytics will give you the results of what is currently happening, so if you did something right, something like Time on Site or conversion should increase for the segment that was served a specialized landing page. Or if your target is to increase regular monthly visitor engagement – look at the number of visitors that return monthly, and see if that is growing.

The quest to fight the swiss cheese product is a battle we all fight every day. In the trenches, it is our job as UX designers to lead the way for the experience to both satisfy the user needs, and make sure the business is getting results where it wants them, By using this technique, you not only produce a list of build able features, you are also making improvements in the areas you need them in a more objective fashion, thus making the overall experience tighter and more effective.