Loading...
Answers
MenuI work for a web design agency. We are trying to make decisions based on data. How do I know which user tests to run and when?
We are a e-commerce web agency. We are currently working on a Startups website and we are driving our design with data. We want to know what tests to run when, how to run them and what deliverables should we send over to our client.
Answers
Sorry to be a bit blunt in my answer, but the client hired your "e-commerce web agency" to do e-commerce conversion optimization, yet your company doesn't know how to do e-commerce conversion optimization?
Perhaps have a conversation with someone in your company or talk to your manager or the agency principal and encourage them to hire for e-commerce conversion optimization expertise. It's an entire discipline requiring years of experience and track record in driving higher conversions for e-commerce sites (in my opinion).
In lieu of that, how many marketing dollars are going to be spent to drive traffic to the site and who is managing that part? Often new start-ups will spend so much of their dollars BUILDING the site, they don't leave much money to spend on AdWords or other methods to drive traffic to it. The amount of traffic will help determine how many tests to run and how often. Usually we start with two very different landing pages (do you have landing pages?), then fine-tune the split testing down to things like headlines and photos. The checkout pages, especially any pricing pages, are also critical in the testing. Is it a brand new concept or established product/service? If brand new, it will take very well-crafted pages to get visitors to pay money for something (vs. a trial). If established, why is your product best and why should I buy now? You should map out the conversion flows to the confirmation page (success page) and measure the drop-off or conversion rates at each step. Test different elements/pages at each step. Also measure things like average cart size (if more than one product), revenue per transaction and long-term customer value.
Hope that helps. Best wishes to you!
Depending if the site has already been built and is going through a redesign VS building a brand new site is very different processes.
Joe points out some critical starting point questions thaty you should have clear answers for in order for you to design with the user in mind.
SITE DESIGN
Much of what we're talking about is building UI design from usability and information architecture methodologies to create the site.
SITE PERFORMANCE OPTIMIZATION
The next step to optimization is to look at site performance and determining if there are areas of improvement which can improve your funnel or optimize your sales and test for varying designs (based on your hypothesis) and use analytics to answer your questions.
MARKETING OPTIMIZATION
Lastly, it's looking into the marketing optimizations where you evaluate marketing ROI, testing of landing pages, and evaluate various marketing channels.
Experts above all offer great information on what to pay attention to. But as mentioned, without details, we are all just making assumptions.
Feel free to schedule a call to discuss further.
Related Questions
-
Will using both Google Adwords Auto-Tagging AND Manual Tagging with UTM variables create any tracking problems in Google Analytics?
I've done a lot of complex analytics installations, usually cross-domain and including AdWords and manual tagging with conversion tracking of ecommerce transactions. The gclid by default will override the manual UTM values in Analytics so that is what you are looking for. If you are just grabbing the manual tags from the URL for your own purpose, this should be OK. However, if you at some point want to import conversions (say) connected to your manual tag, then you would want to enable UTM overriding the gclid: https://support.google.com/analytics/answer/1033981?hl=en. Do you have AdWords and Analytics accounts linked? Where and how are you tracking commerce conversions, if you are? Do you feel you are getting the insights you need from your current custom reporting? Let me know if I can help with a call about those or further optimization of your analytics capture. Best - MichaelML
-
How do I run a closed beta test for my mobile application? Development will be finished in 3 weeks.
You should try to engage people using social networks, it is easier to spread than email. The conversion rate on emails are low but is still a valid tool for that. Send and email with a simple and objective message that will make people want to try. The best way to have feedback from users is to watch them use the app. You should put them on the hands of everybody that you can and without any instrucions and just watch, don´t even say that the app is yours. Try to do it a lot. If you want feedback from others, you can include the feedback form inside the app and suggest users to answer occasionally. I would also strongly recommend to use a tool as Flurry Analytics. Is the best way to get data from how is the use of the application. Pay attention to those data and be open to change your app a lot, you may need more features or cut some off to make it easier to use. If you need more help please contact me.BS
-
How should the dynamic between a ux designer and a developer who are working together look like?
It depends a lot of in the skill sets and experience of both people but in most cases the ux designer should be controlling the developer pretty heavily in order to make sure his ideas come through properly. The UX designer may just need to work on his approach so people don't feel bossed around and more like they are working together. In an ideal world, there would be a project manager who makes sure everything is communicated well and keeps the dynamic feeling great.JM
-
Is having "HOME" button in navigation menu necessary if I have a clickable logo? What makes most sense from UX POV?
We have been collecting usage data on the home button from about 750 websites we manage across North America in an effort to try to determine if it is necessary or not. While each website is different, and much of the data is statistically insignificant, we have started to operate with a few assumptions. 1.) Most users, particularly younger users, recognize the logo as a way to get to the home page. 2.) Websites without home buttons seem to get a comparable amount of traffic to the home page as those that do not. We don't see a significant difference between having a home button or not. 3.) Websites without a home button often will see an increase in direct traffic from returning users during a session indicating that users who do not know the logo is a route to the home page will instead clear the address bar back to the root domain to get back home. Based on our research, we have decided to omit the home button in most instances. Although when it is present, it is often used, most users seem to understand how to get to the home page regardless of the inclusion of a home button. With the complexity of modern websites, we are usually pressed for space in the header and can better use the real estate that would be dedicated to the home button for other UI elements. That said, if the audience for our website skews older, we will still include the home button. Our research has indicated older users are less familiar with the concept of the logo being a home button.RS
-
What's your opinion on using something like usertesting.com vs. real time usability testing (online and offline)?
UserTesting can be instructive in terms of understanding whether people understand your copy, CTAs, and intended flows but generally, I've found the quality of their panels to be pretty low. You're almost always getting people who are not your actual users, so the feedback can only be generally applied as above. I find whatever web analytics package or packages you're using are generally able to provide much better insights. I also really do believe in *real* user panels. Buying pizza or offering small financial incentives to real users to click through new flows where they are talking out loud or answering specific questions is going to give far more actionable insights than anything else. What I like to do is take my best guesses as to what's not working or what I'm looking to improve and then discover/validate via real in-person customer panels. Happy to talk through this in more detail with you in a call.TW
the startups.com platform
Copyright © 2025 Startups.com. All rights reserved.