<?xml version="1.0" encoding="UTF-8"?><rss xmlns:dc="http://purl.org/dc/elements/1.1/" xmlns:content="http://purl.org/rss/1.0/modules/content/" xmlns:atom="http://www.w3.org/2005/Atom" version="2.0" xmlns:cc="http://cyber.law.harvard.edu/rss/creativeCommonsRssModule.html">
    <channel>
        <title><![CDATA[Stories by María Barrena on Medium]]></title>
        <description><![CDATA[Stories by María Barrena on Medium]]></description>
        <link>https://medium.com/@mbarrena?source=rss-57049ad8ac6b------2</link>
        <image>
            <url>https://cdn-images-1.medium.com/fit/c/150/150/1*Q31cO2cEn8ssNdMunnjAGA.jpeg</url>
            <title>Stories by María Barrena on Medium</title>
            <link>https://medium.com/@mbarrena?source=rss-57049ad8ac6b------2</link>
        </image>
        <generator>Medium</generator>
        <lastBuildDate>Wed, 18 Nov 2026 17:39:21 GMT</lastBuildDate>
        <atom:link href="https://medium.com/@mbarrena/feed" rel="self" type="application/rss+xml"/>
        <webMaster><![CDATA[yourfriends@medium.com]]></webMaster>
        <atom:link href="http://medium.superfeedr.com/" rel="hub"/>
        <item>
            <title><![CDATA[Evaluating GenAI features in enterprise software requires a human touch]]></title>
            <link>https://medium.com/@mbarrena/evaluating-genai-features-in-enterprise-software-requires-a-human-touch-9ab4a7f0fabf?source=rss-57049ad8ac6b------2</link>
            <guid isPermaLink="false">https://medium.com/p/9ab4a7f0fabf</guid>
            <category><![CDATA[gen-ai-tools]]></category>
            <category><![CDATA[user-experience]]></category>
            <category><![CDATA[enterprise-software]]></category>
            <category><![CDATA[metrics]]></category>
            <dc:creator><![CDATA[María Barrena]]></dc:creator>
            <pubDate>Wed, 08 Apr 2026 09:38:16 GMT</pubDate>
            <atom:updated>2026-04-08T09:38:37.058Z</atom:updated>
            <content:encoded><![CDATA[<h4>How to assess accuracy, time savings and errors to further improve the quality of responses</h4><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*v-tJjET6gY9yVonbbvxFXQ.jpeg" /><figcaption>Technician Examining Computer Hardware Components. <a href="https://www.pexels.com/photo/technician-examining-computer-hardware-components-31725401/">Photo by Bulat843 from Pexels</a></figcaption></figure><p>Last year, a great working opportunity came knocking at my door: I was going to lead the design process of the (and my) very first GenAI-powered feature on an enterprise software tool. A feature that may allow users to set up and onboard their content in it faster.</p><p>Back in 2024 and 2025 companies spent some time at first assessing whether the technology was good enough for their customer’s needs and we were not any different. By the time I got involved, there was a working prototype, hardcoded and designed by engineers, that required a lot of love and polishing. No UX designer was involved.</p><p>Even though improving the user interaction was a must, understanding the quality of the generated responses by the LLM behind the scenes was priority number one. Moreover, any feature that relies on AI under the hood is often perceived as a productivity booster tool, despite <a href="https://skooloflife.medium.com/ai-isnt-making-you-productive-it-s-making-you-feel-productive-37268d8c5c98">voices in the industry stating that AI does not really improve productivity</a>. So how much time, if any, were users going to save by benefiting from it?</p><p>This article describes a way to assess accuracy, time savings and errors based on my experience and learnings working on that feature.</p><h4>Subject-matter expert evaluation</h4><p>There are multiple ways of measuring how <strong>accurate</strong> the responses (or outputs) GenAI generates are. You may have heard about the LLM-as-a-judge technique where an AI measures how good AI responses given are, in a AI-evaluates-AI way. And that’s a great technique when:</p><ul><li>Outputs are always the same for the same piece of input. For example, if you ask any LLM <em>when is St. Patricks day?</em> there is only one correct response: March 17th.</li><li>Outputs may differ with the same input but you have TONS of data and rules that can be used to train the LLM in analysing the responses. For instance: a medical app that detects a tumor by analysing a medical scan because the trained data consist on millions of medical scans.</li></ul><p>When your database is minimal, your output subjective or don’t have access to public or copyright-free data to train your system, the situation of any enterprise software out there, we need to consider human evaluation.</p><p><strong>Subject matter experts</strong> are vital to assess how good your Gen-AI tool is doing the analysis and providing responses. Like a Technician who can better assess what’s wrong with your device, experts are the ones who can look for nuances the AI might miss.</p><h4>Evaluating search queries as inspiration</h4><p>Great! We have the <em>who</em> already. Let’s figure the <em>how</em> now.</p><p>If you have been alive between the end of 2023 and whatever year is it when you are reading this, you are very aware of how much AI has evolved. I can confirm that as in January 2025, the literature related to measuring quality of responses was minimal so we had to expand our horizons.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*64k7f4pwUoZDU7M49XfAAA.jpeg" /><figcaption>A Person Shopping Online. <a href="https://www.pexels.com/photo/a-person-shopping-online-5585790/">Photo by Polina Tankilevitch from Pexels</a></figcaption></figure><p>I started looking into how to measure quality of search query results instead, especially in eCommerce. An industry I am very familiar with, where results may not be what users expect to begin with. The task seem close enough to our feature and the amount of papers and trust worthy online resources was way wider.</p><p>I came across two interesting methods: the <a href="https://docs.smith.langchain.com/evaluation/concepts">Langsmith evaluation</a> and the <a href="https://assets.amazon.science/3d/9d/6a8e1d294c2d80ac955eb00fcbae/beyond-accurate-answers-evaluating-open-domain-question-answering-in-enterprise-search.pdf">Amazon’s Search Answer Quality Type</a>. While in the first one, evaluators need to rank each task performed separately, amazon’s method ranked the overall quality of answers.</p><p>The output of the GenAI feature consisted in multiple data points, all derived from one single task. In cases like these, it is relevant to measure the quality of each data point separately, to understand which piece needed to be improved post measurements. <strong>Merging these two methods</strong> into one is my chosen approach.</p><p>In summary: to measure accuracy, all we need to do is letting subject matter experts try the GenAI feature out and rate each data point, giving each an answer weight depending on their level of importance.</p><p>The summary of responses per participant becomes then our percentage of response accuracy. This is how it looks:</p><blockquote>Q: The amount and kinds of suggested <em>#datapoint1</em> are:</blockquote><blockquote>A: PERFECT (100%), TO-BE-VERIFIED(75), INCOMPLETE(50), PLAUSIBLE(25), IRRELEVANT(0), NQ (NQ)</blockquote><blockquote>— Answer weight 40%</blockquote><blockquote>Q: The amount and kinds of suggested <em>#datapoint2</em> are:</blockquote><blockquote>A: PERFECT (100%), TO-BE-VERIFIED(75), INCOMPLETE(50), PLAUSIBLE(25), IRRELEVANT(0), NQ (NQ)</blockquote><blockquote>— Answer weight 40%</blockquote><blockquote>Q: The amount and kinds of suggested <em>#datapoint3</em> are:</blockquote><blockquote>A: PERFECT (100%), TO-BE-VERIFIED(75), INCOMPLETE(50), PLAUSIBLE(25), IRRELEVANT(0), NQ (NQ)</blockquote><blockquote>— Answer weight 20%</blockquote><p>These responses will set the direction: it will hint you the areas to improve by applying prompt engineering afterwards.</p><h4>Benchmarking time on task</h4><p>Since we are going to allow experts to test the feature, we can ask them to first perform the task without leveraging the new feature and measuring the time on task. Afterwards, they complete the same task by using the AI-powered feature, measuring the time invested on it again.</p><p>By collecting an initial metric without the tool and then with it, we <strong>benchmark</strong> the time on task. We compare those two sets of times to measure time savings. Furthermore, measuring time savings provides a clear value metric for stakeholders. It transforms “AI magic” into a measurable business benefit.</p><h4>Beyond task efficiency</h4><p>Due to its Gen-AI nature, we are putting the focus on measuring Accuracy and Task efficiency for this feature. However, in enterprise software, “mostly correct” is rarely sufficient. And that means we should also pay attention to Errors.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*UuBddFz2eG6Rem7iZp5wmg.jpeg" /><figcaption>An error in the console. <a href="https://www.pexels.com/photo/creative-internet-typing-abstract-6212801/">Photo by Markus Spiske</a></figcaption></figure><p>One way to measure errors in Gen-AI features is by counting the amount of hallucinations participants experience when interacting with the feature and what Data Point of those identified in the <em>accuracy</em> metric is affected by it. These two data sets can take you to polish the underlying instructions powering the feature.</p><p>Errors and other user experience aspects such as cognitive load or learnability are worth to measure in enterprise software features. For that, the <a href="https://www.nngroup.com/articles/castle-framework/">CASTLE framework</a>, The Nielsen Norman Group’s response to the HEART framework by Google, has proven to be a great way to establish and decide on the metrics on workspace applications, to assess tools and features, whether they are powered by AI or not.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=9ab4a7f0fabf" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Applying UX copy techniques to name an API]]></title>
            <link>https://medium.com/@mbarrena/applying-ux-copy-techniques-to-name-an-api-eb3dace9926a?source=rss-57049ad8ac6b------2</link>
            <guid isPermaLink="false">https://medium.com/p/eb3dace9926a</guid>
            <category><![CDATA[ux]]></category>
            <category><![CDATA[api]]></category>
            <category><![CDATA[enterprise-ux]]></category>
            <category><![CDATA[research]]></category>
            <dc:creator><![CDATA[María Barrena]]></dc:creator>
            <pubDate>Fri, 08 Oct 2021 10:00:42 GMT</pubDate>
            <atom:updated>2023-04-01T18:14:30.449Z</atom:updated>
            <content:encoded><![CDATA[<h4>How did we experiment with a UX copy assessment to get the right name for a new resource in our platform</h4><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*D8y2CpUxN3YrlxS5AF-CKA.jpeg" /><figcaption>Photo by <a href="https://unsplash.com/@sure_mp?utm_source=unsplash&amp;utm_medium=referral&amp;utm_content=creditCopyText">Surendran MP</a> on <a href="https://unsplash.com/?utm_source=unsplash&amp;utm_medium=referral&amp;utm_content=creditCopyText">Unsplash</a></figcaption></figure><p>At my team, we love experimenting and trying out innovative ideas that let us get the best possible solution for our users. This time we used an existing framework for an unusual purpose: to come up with a name for a new feature and its resource in our platform.</p><p>But before spoiling all the fun, let’s go back to the initial problem we were trying to solve. Based on previous research conducted in our eCommerce company, we were aware that some customers selling products in different storefronts were often creating one project per storefront needed in our product. As a consequence they end up duplicating data and/or synchronising it between projects.</p><p>One of the reasons for that setup was that each storefront had to sell slightly different product catalogs. Imagine having a product catalog of 500K products, all with their product data and pictures, where the 95% of it is available in all storefronts the customer has. We needed to improve our customer experience so they could avoid the multi-project setup scenario and just send the right data to each storefront, or in case of B2B, to each customer.</p><h4><strong>Call me by your name</strong></h4><p>During the discovery phase of the project, when the team was gathering use cases to <strong>define the problem</strong> better and model the domain, we conducted multiple interviews with targeted users of our APIs. It’s during these initial steps when the team usually establishes the <a href="https://martinfowler.com/bliki/UbiquitousLanguage.html#:~:text=Ubiquitous%20Language%20is%20the%20term,language%20between%20developers%20and%20users.&amp;text=Evans%20makes%20clear%20that%20using,and%20hence%20the%20domain%20model."><em>Ubiquitous Language</em></a>, a common language between team members and our users.</p><p>However, we realised that <strong>different participants were using a variety of terms to designate the same type of functionality</strong>: a group of products available on a particular storefront. The names were: assortments, catalogs or collections to name a few. This often resulted in the interviewer, one of our commercetools folks using one term and the interviewee, an end-user (usually an engineer), answering by using a completely different word.</p><p>Based on that experience, we realised we could just not ignore it and go for the initial name we had in mind. Looking at the official definition of each term and the names already used in the industry for similar solutions made the list of options even longer, adding more terms to take into consideration.</p><blockquote>Instead of just talking to people in the business as we were already doing, we decided to empirically find out what users would call the functionality based on the capabilities we were going to offer.</blockquote><p>That’s when we decided to follow the Domain-Driven Design approach suggested by Eric Evans in his book <a href="https://www.oreilly.com/library/view/domain-driven-design-tackling/0321125215/">“Domain-Driven Design: Tackling Complexity in the Heart of Software”</a>, but instead of just talking to people in the business as we were already doing, we decided to empirically find out what users would call the functionality based on the capabilities we were going to offer. In that way we could make sure that it is easy to discover when using our platform and unlike Elio, the protagonist of the novel <a href="https://www.goodreads.com/book/show/36336078-call-me-by-your-name"><em>Call me by your name</em></a>, we were going to <em>“have the courage to ask a question like that.”</em></p><h4><strong>Introducing an altered version of the cloze test</strong></h4><p>If you have ever learnt a second language in school, you are definitely going to be familiar with the test we conducted. Remember those exercises in your workbook where you had to fill in the gaps in a paragraph? That’s what’s officially called a cloze test.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*YoKrGE0TkIdjqMznZAcuPw.jpeg" /><figcaption>A cloze test in a German workbook</figcaption></figure><p>The cloze test was <a href="https://www.gwern.net/docs/psychology/writing/1953-taylor.pdf">created by Wilson L. Taylor in a workshop in the 50’s</a> as a way to measure readability. However, the framework has been used for multiple purposes throughout the time:</p><ul><li>By teachers as mentioned above, to reinforce vocabulary, grammatical knowledge and reflects linguistic competence of a language by students,</li><li>By UX or tech writers, to assess readability and comprehension of some kind of content and verify the content suits the targeted audience. This was really well explained by Sven Heckler in his wonderful article <a href="https://techblog.commercetools.com/bring-your-apis-from-good-to-awesome-with-ux-research-15980ad18d24">Bring your APIs from good to awesome with UX research</a>.</li></ul><p>The theory behind it says that when we are in front of a cloze test, we combine our <strong>background knowledge, previous experiences and our reasoning</strong> in order to fill in the gaps.</p><blockquote>Instead of verifying that “the content suits the audience you are targeting”, we were interested in finding <strong>the most meaningful name for the new functionality and resource in our platform for all types of users</strong>. That was the trigger for our experiment and the reason why we considered using an altered version of the cloze test as the means for reaching our goal.</blockquote><p>Given the purpose of it and to distinguish it from previous Cloze tests, we are going to call it the <strong><em>Denominative Cloze Test.</em></strong></p><h4><strong>Preparing a Denominative Cloze Test: </strong>the step-by-step guide</h4><p>You don’t have to be a highly skilled researcher to conduct this study. In fact, if you are a developer building your own API and have good writing skills, you can easily run the experiment. Here are the steps to follow:</p><ol><li><strong>Create a 125–200 word paragraph</strong> defining what users can do with the new resource, like the ones you usually find in technical documentation, as in this case your target audience are technical users. Use one of the options you have in mind throughout the paragraph. Then, substitute that resource name you were using with blanks, one for each time it appears on the text. <br><br>The following paragraph is the beginning of our Denominative cloze test: <br><br><em>With the new resource called _____________, merchants are now able to group a specific subset of products to make them available in their separate storefronts or output channels. Once you create a ___________ in your project, there are two ways to…</em></li><li>Write down also a <strong>brief introduction</strong> of the task to participants. The rules for the introduction are:<br>• Try to explain briefly the new functionality but do not use any of the possible terms you are considering.<br>• Explain the purpose of the task and the reason behind it: we realised that when we did not explain the purpose of the test, some participants assumed we were being asked for their opinion on readability of the content. By being truthful to our intentions, users were more likely to give an open response, making them reasoning on their answers. <br><br>If you are a researcher or a UX writer reading this article you might consider this usage of the cloze test controversial, as on a regular test where you are testing readability, the paragraphs tend to be longer (the longer the texts are, <a href="https://www.researchgate.net/publication/314794141_What_Is_the_Optimum_Length_of_a_Cloze_Test">the better results you get</a>) and words behind the gaps will refer to different terms. But remember we are using this framework for a different purpose: <strong>to find out the best resource name.</strong></li><li><strong>Find participants:</strong> the best possible participant type are end-users, developers familiar with your or similar APIs. When end-users are not available for the test, consider people part of your organisation with certain domain knowledge such as tech writers or those who are in contact with your end-users often. If the session is conducted on a call or in person, schedule a 20 min session with each of them separately. We conducted a total of 11 sessions and in our case, they gave us meaningful insights, enough to take a decision.</li></ol><blockquote>By being truthful to our intentions, users were more likely to give an open response, making them reasoning on their answers.</blockquote><h4>How to conduct one</h4><p>Before conducting any session, I encourage you to <strong>test the test with a colleague.</strong> There are always last minute tweaks on the paragraph that you might want to do. Changing the paragraph in the middle of the research study may lead to inconclusive results as knowing how much the change you did has an impact on the participant’s suggestions is hard to identify.</p><p><strong><em>Denominative Cloze Tests</em></strong> can be conducted in two ways:</p><p><strong>Option 1: On a video call or in person<br></strong>The session happens synchronously as you, the tester, will be present and can bring a note taker. <a href="https://uxplanet.org/implications-of-hawthorne-effect-on-users-behavioral-psychology-5d41d6760ea4">Stay away from the Hawthorne effect</a> by not inviting too many people to the call. <br>The session is divided in the following way:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*85qy16dXcH6tIZCd0830qg.jpeg" /><figcaption>Photo by <a href="https://www.pexels.com/@greta-hoffman?utm_content=attributionCopyText&amp;utm_medium=referral&amp;utm_source=pexels">Greta Hoffman</a> from <a href="https://www.pexels.com/photo/crop-person-taking-notes-while-working-7674917/?utm_content=attributionCopyText&amp;utm_medium=referral&amp;utm_source=pexels">Pexels</a></figcaption></figure><ul><li>Start thanking the participant for their time</li><li>Read them the introduction you wrote and invite the participant to think aloud to hear their thoughts throughout the session</li><li>Share the paragraph with blanks with them in any way you prefer: on the chat if you are on a video call, by sharing it on your screen, printed on a paper if in person, etc.</li><li><strong>Let participants speak and listen to their reasoning.</strong> If you have not invited a colleague taking the role of note-taker to your sessions, take notes of their preferred term(s) and specially, write down the reasons why they have chosen it.</li><li>Naturally, <strong>stay neutral throughout the process</strong> and avoid giving any hints to participants of your preferred term. Otherwise you’ll be just conducting this test with the purpose of validating your personal opinion and that should not be your goal.</li><li><strong>Embrace the silences </strong>while the participant goes through the paragraph and their thoughts. Trying to fill in the blanks requires a certain level of cognitive effort and they might need some time to think about their answer.</li><li>Once the participant has finished their reasoning, ask them for their opinion on other terms you have considered and why/why not they would choose them.</li></ul><blockquote>Unlike typical cloze tests, <strong><em>Denominative cloze tests</em></strong> don’t have <em>wrong</em> answers.</blockquote><p><strong>Option 2: Through a survey<br></strong>If you cannot conduct the session simultaneously or have a time constraint, you can set up a survey and collect feedback asynchronously. The survey should contain both the introduction and the paragraph with blanks. Then add two field for users to provide both their preferred term and the explanation why they went for that option.</p><p>As an optional step, you could provide a second part in your form, including some of the terms you considered and asking participants what they evoke for them when they are used.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*Z1lxtOAzYraelzWMTsbOrQ.jpeg" /><figcaption>Photo by <a href="https://www.pexels.com/@thisisengineering?utm_content=attributionCopyText&amp;utm_medium=referral&amp;utm_source=pexels">ThisIsEngineering</a> from <a href="https://www.pexels.com/photo/female-engineer-writing-on-whiteboard-3862380/?utm_content=attributionCopyText&amp;utm_medium=referral&amp;utm_source=pexels">Pexels</a></figcaption></figure><h4><strong>Sharing the results</strong></h4><p>Once you have conducted all the sessions, gather all the terms you were given and list all the reasons stated by participants. Include the reasons against the term as well if you got any. In our case, they drove our decision.</p><p>Then <strong>share the results with your team</strong> and with all that evidence open up for discussion.</p><blockquote>Discussing and analysing the user’s reasons should guide the team on finally deciding on a term which may or may not be one of the options given by participants.</blockquote><h4><strong>Taking the final decision</strong></h4><p>Coming back to our initial task, in our case, demographics and industry where participants worked definitely played a relevant role in their chosen options. Having a large target audience was the reason why we had a handful of different terms even before launching our experiment, so the bigger your target audience is, the more meaningful running a <strong><em>Nominative Cloze Test</em></strong> could be for you.</p><p>In our case, we did not go for the most chosen term, <em>catalogs,</em> as the reasons against using it were strong enough to dismiss it: too traditional, less innovative and very tied to the retail industry. A similar case happened with <em>assortments </em>and other options that were too connected to a certain user group but sounded clearly foreign to others. Once we had reviewed them all, we decided to go for a variation of one of the options a participant gave us: <em>Product Selections</em>.</p><p>Product Selections was <strong>generic and neutral enough to be understandable by all audiences</strong> and did not have any particular meaning previously tied to it as terms more specialised and linked to a certain industry had.</p><p>Even though we chose a term that was not suggested multiple times, the <strong><em>Denominative Cloze Test</em></strong> has guided us throughout the process. Without investing the time in conducting this little research experiment, we would have never come up with a term that, from then on, software developers, domain experts and even end-users use.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=eb3dace9926a" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[How we transformed the product search for the better]]></title>
            <link>https://medium.com/@mbarrena/how-we-transformed-the-product-search-for-the-better-7f8ac75dbeba?source=rss-57049ad8ac6b------2</link>
            <guid isPermaLink="false">https://medium.com/p/7f8ac75dbeba</guid>
            <category><![CDATA[search]]></category>
            <category><![CDATA[ux]]></category>
            <category><![CDATA[filters]]></category>
            <category><![CDATA[case-study]]></category>
            <category><![CDATA[ui]]></category>
            <dc:creator><![CDATA[María Barrena]]></dc:creator>
            <pubDate>Tue, 11 Feb 2020 16:17:38 GMT</pubDate>
            <atom:updated>2023-04-01T18:00:41.974Z</atom:updated>
            <content:encoded><![CDATA[<h4>A UX case study of how we built our new search and filter solution for eCommerce product catalog managers, by following a user-centered design approach</h4><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*V938Uny0ycE8cGhO_zKqGg.png" /><figcaption>The new Merchant Center product search and filter visual interface</figcaption></figure><p>The product search functionality available in the Merchant Center, an enterprise software application, lets merchants find products in their product catalog.</p><p>Originally, we had a single search implementation catering to both catalog managers and shop customers. Over time, we noticed that their characteristics differed: storefront search queries are simpler and returning results super quickly is crucial. However, merchant searches are more complex. And optimizing for complexity <strong>and</strong> speed is complicated.</p><p>By using a brand new feature as an example, we would like to show how the product team follows a user-centered design approach to provide the best solution for all kinds of needs.</p><h3><strong>The challenges</strong></h3><p>Despite sharing the same goal of finding products, the two groups of users have different needs and intentions. The principal end-goal for storefront customers is to make a buying decision. Merchants can customize the storefront search to promote certain products. But their principal need in a catalog search is to find products to check the data quality or update it, publish new products or manage inventory.</p><blockquote>The merchants principal need in a catalog search is to find products to check the data quality, publish new products or manage inventory.</blockquote><p>The accuracy of results returned in the Merchant Center catalog search was also a concern from some merchants. This was not ideal for them, whose more complex search needs also led to slow performance and a decrease of productivity.</p><h3><strong>UX and UI goals</strong></h3><p>The existing visual interface in the product section of our Merchant Center was not suited for our heavy users (those that use it to a great extent). Some users even admitted they have never tried the built-in filters. For those using them, filtering by attribute data was especially cumbersome as they had to:</p><ol><li>Open a pop-up modal</li><li>Find the filter attribute</li><li>Search for or type the desired value</li><li>Apply and close the modal</li></ol><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*HE8suqZIqt5_aUdMnQr2UA.gif" /><figcaption><em>The more attributes a catalog contained, the harder the filter is to find.</em></figcaption></figure><p>In the most common screen resolution, 720p (1280x720 pixels), the results list would even move below the fold after applying a fourth filter. Definitely, the new filters needed to be more accessible.</p><p>Catalog managers with complex data models need to view loads of information in the product list. But those with simpler data models don’t need much data displayed in the product list. For them, the search and filter functionalities are more relevant. By <strong>decluttering the page and making it more customizable </strong>we can accommodate those two different usage patterns. This lets Merchant Center users focus on what’s relevant for them in the new list.</p><h3><strong>Research and initial solutions</strong></h3><p>Let’s say we have a product catalog containing a product called “Perfect Blue DVD” and another one called “Nike shoes”. The latter contains a searchable attribute “color” with the value “Blue” selected. With the existing implementation, both products will show up in the results when a catalog manager types the term “Blue” in the search input.</p><p>This behaviour might match the storefront customer’s intention. But the merchant doesn’t need to find all products that contain the term “Blue” on any field of their product information. Ideally, <strong>they would target which part of the product data they would like to look into</strong> (the product name OR the color field). This was the reason why some catalog managers reported inaccuracy of results.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*Vh5l8eJvT9DmOiQ3Lb0abg.png" /><figcaption>Result list showing products from a brand called “Blue” among the results</figcaption></figure><p>Decoupling storefront and merchant search into two separate entities solves the problem: keep the existing search API for the storefront search and create a new search adapted for merchants needs. That would mean merchants could perform their tasks without affecting the storefront search.</p><p>By splitting the searches, we could now benefit from <strong>faceted search.</strong> The main search input would be the place to search for basic product data (names, descriptions or SKUs). For any other characteristic, we would provide with more advanced filters for each one of their attributes. For instance, if a catalog manager is looking for black accessories, the facets used are one in categories and another one in the attribute color.</p><p>This solution brought usability questions:</p><ol><li>How do we help users leverage the power of filters when searching for attribute values?</li><li>How do we encourage them to start using the filters again?</li><li>What’s more relevant for merchants, see their entire product list or have filters always visible and available? The use cases might differ from one company’s data model to another. A way for each merchant to choose among those two options would be the best solution.</li></ol><h3><strong>Prototypes and wireframes</strong></h3><p>The decision whether to move the filter section (and if we did, where) was one of the first UX/UI decisions taken into the project. Filters work the same way and are always placed in the same area of the page throughout the whole Merchant Center platform. Moving them to a different place on this single page would solve some problems but it would break the consistency in our UI. Oh, the old battle between consistency (one of <a href="https://www.nngroup.com/articles/ten-usability-heuristics/">the 10 usability heuristics of user interface design</a>) and efficiency that UX designers have been fighting many times!</p><p>A common design pattern for filtering a list is a<strong> filter sidebar.</strong> Many online shops and travel websites (e.g, Amazon or booking.com) use that UI solution to accommodate their large number of filters (catalog manager’s situation as well!). The benefits we see implementing a sidebar in the product list for our particular case are:</p><ul><li>Have all filters in a single place (no more pop-ups needed)</li><li>The applied filters won’t push results below the fold</li></ul><p>To cover the needs of every merchant, and by taking into consideration the <a href="https://www.interaction-design.org/literature/topics/progressive-disclosure">progressive disclosure pattern,</a> we came up with the idea of showing and hiding the filter sidebar on demand. The sidebar is hidden by default, so it does not overwhelm the user. When users don’t need to filter products by attributes, they can hide it and focus on their product list.</p><p>As Katie Sherwin describes in her article <a href="https://www.nngroup.com/articles/applying-filters/"><em>User intent affects filter design,</em></a> there are two dimensions to take into consideration during the design of filters: when to apply the filters to the result list and when to scroll to the top of the result list.</p><blockquote>There are two dimensions to take into consideration during the design of filters: when to apply the filters to the result list and when to scroll to the top of the result list.</blockquote><p>If the user’s intention is to start scanning the result list right after applying a filter, refresh the result list and scroll back to the top provides the most useful experience to the user. But what if the user is still browsing and applying filters? Taking them back to the top of the page would force them to scan the sidebar again, to find where they were before the disruption. We needed a way to keep the filter sidebar untouched when the product list refreshes. Independent scrolling was the answer.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*wjipJ9LnkhPkB43j65Ez0w.png" /><figcaption>A filter sidebar with independent scrolling allows users to put focus on their current task.</figcaption></figure><p>Once we figured out the page structure, it was time to focus on the sidebar content. Some stores have huge data models with hundreds of product attributes. Showing all these attributes in the filter sidebar by default would not be too useful for them, as the sidebar will become too cluttered and filters hard to find. They needed a way to edit the sidebar: choose what filters they need in that sidebar and remove what they don’t need.</p><p>The first sketches included a filter sidebar with two tabs:</p><ul><li>We would put the basic filters in the default tab: those that any product catalog would contain.</li><li>The second tab would contain the attribute filters (color, size, brand…), which were previously hidden in the pop-up.</li></ul><p>By initial interviews and shadowing some Merchant Center users, we found out that each catalog manager uses different types of data to find their products. Thus, we assumed first and by user testing then confirmed that the distinction was not even needed. Putting all filter attributes together in one place will improve their discoverability.</p><figure><img alt="different states of development of the merchant center filter sidebar" src="https://cdn-images-1.medium.com/max/1024/1*N_cJQcwakxR54QhdTGsTWw.png" /><figcaption>The first wireframe (left) containing the filter sidebar with two different tabs and how the sidebar concept changed throughout the research process.</figcaption></figure><p>Each type of filter required a different user interaction depending on the type of data merchants were looking for:</p><ul><li>Filters where the data is a list of items could be easily applied to the list by clicking on the item.</li><li>Filters that allow any text or numeric value would need an input where merchants could submit any value and press enter to submit the filter.</li><li>Filters for date and/or time ranges would contain two inputs to submit either a single value or a range.</li></ul><p>In some cases such ranges, submitting by pressing enter might not be intuitive for those less tech savvy users. We added a button for better usability instead.</p><p>Despite the difference, we needed a standardized, easy way to apply and remove several filters from the list. Let’s imagine that a merchant needs to find all products from the category “accessories” created last month and today. This merchant would need to search for a date range and for a specific date in the “creation date” filter. The solution was to display any applied filter with a checkbox, allowing users to apply and remove each filter with one click.</p><p>Last but not least, we wanted to take advantage of the information that facets provide. Airbnb’s use of facets in their price filter makes for a good example and source of inspiration for us. They use them to visually display the amount of places available by price range. This information can give Airbnb users a sense of what’s the average price in a location. But let’s not forget that these users are looking for one single item (in this case, accommodation), while the merchants need is quite different as explained before. Showing the average number of products per filter is not that relevant for their usual tasks.</p><figure><img alt="price filter airbnb" src="https://cdn-images-1.medium.com/max/311/1*fvZrIQA2H3TKw1Xch1bmPg.png" /><figcaption><em>Airbnb price filter</em></figcaption></figure><p>A solution found often in e-commerce is the use of numbers that represent the amount of products containing that filtered value. With this visual hint, merchants don’t even need to apply the filter to know how many products they will get. With this usage of facets, we follow the <a href="https://books.google.de/books?id=heE_AAAAQBAJ&amp;pg=PA76&amp;lpg=PA76&amp;dq=search+results+poka+yoke&amp;source=bl&amp;ots=fVLLMSjpFE&amp;sig=ACfU3U2YKTf3X89OWdng9FNzvNkiRWTzAw&amp;hl=en&amp;sa=X&amp;ved=2ahUKEwjOwt2utsfnAhULUBUIHUdFDcsQ6AEwGnoECAoQAQ#v=onepage&amp;q&amp;f=false">Poka-yoke principle</a> and prevent them from getting zero results.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/364/1*KYkwPhdxIaGVuqKzCWqQIQ.png" /><figcaption>Suggested values represented as checkboxes, ready to be applied as filters</figcaption></figure><p>Depending on the merchant’s need, suggesting values into each filter could be helpful. But for now, the only data we could get was the value included into the biggest amount of products. Without applying machine learning on each project, these suggestions might not be useful for merchants. As this thought was only an assumption, We still wanted to get feedback from real users about it.</p><p>We decided to check if we were on the right path before implementation as the change in both page structure and functionality were too broad. The most effective way to do so is by <strong>testing</strong> prototypes with real users.</p><h3><strong>Usability tests and results</strong></h3><p>Five product catalog managers, real users of the Merchant Center, were subjects taking part of the usability tests. There are many studies (like <a href="https://www.nngroup.com/articles/how-many-test-users/">this one</a>) that prove that with five subjects you get close to user testing’s maximum benefit-cost ratio. With me in the role of the facilitator and a fellow UX designer as a note-taker (so we did not miss our user’s reactions), we visited them at their workplace. In order to replicate the user’s usual scenario, the product data we used was part of their own product catalog.</p><p>The 15-minute test consisted of a short interview and a task using a clickable prototype. The three-part task was the following:</p><ul><li>You need to perform an update in all the antifreeze products under the “Kitchen and household” category between 2 and 5 liters.</li><li>You have just realised that you just need to update the antifreeze products that are smaller than 2 as well.</li><li>There are too many filters on your list that you no longer need. Is there any way to solve that?</li></ul><p>We asked participants to express their thoughts while performing the task. This method, called <a href="http://grouplab.cpsc.ucalgary.ca/saul/hci_topics/tcsd-book/chap-1_v-1.html"><em>thinking aloud</em></a>, gives qualitative feedback as the thoughts are shared in real-time.</p><p>The goals of the usability test were the following:</p><ul><li>Learning if the filter sidebar is discoverable</li><li>Verify if the visual representation of facets is understandable</li><li>Follow the actions that users will perform to find products of a specific product type and from a certain attribute value range</li><li>Confirm whether the “edit sidebar” feature was easy to discover and use.</li></ul><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*oSrEcE1Woy9HifpO" /><figcaption>The more low-fidelity your prototype is, the less afraid participants will be in communicating the problems they encounter.</figcaption></figure><p>Our findings after testing the prototype were:</p><ul><li>2 out of 5 test users <strong>tried to use the search input in order to fulfil all the tasks, dismissing the filters.</strong> They typed all the possible values that the task provided and tried to submit the search query.</li><li>1 out of 5 test users mentioned that they rarely need to search for product types or categories. This catalog manager focuses on filter attributes instead. <strong>Not distinguishing between basic filters and attribute filters was a good decision.</strong></li><li><strong>None of the test users found helpful the suggested values on filters.</strong> We decided not to implement that as it would not bring any value for them.</li></ul><h3><strong>Beta release</strong></h3><p>Launching a new interface, especially with a completely new API behind is always risky. Our team decided to release this new functionality as a beta to a select group of users first as well as those who participated in the usability tests. The usability tests provided good input, but we wanted to get feedback on the initial implementation, to learn and improve.</p><p>Concurrently, a different group of UX/UI designers tried to replicate usual catalog managers tasks first and then ranked the usability breaks found as high, medium and low risks. Thanks to those, we were able to improve both copy and usability prior to the open release. This process is called <a href="https://uxdesign.cc/ah-what-could-go-wrong-a-k-a-risk-analysis-for-user-interfaces-part-i-6c3a6973e66">UI risk analysis</a>.</p><blockquote>A different group of UX/UI designers tried to replicate catalog managers tasks first and then ranked the usability breaks found as high, medium and low risks, following a UI risk analysis approach.</blockquote><p>After a few days, we had the chance to go through the new product search with some of these users. With their questions and their feedback, we were able to understand if the new solution fulfil their needs.</p><p>If the users of your interface need to complete some training to start using it, it means that interface is not intuitive enough. We needed a way to highlight the new features in order to encourage all users to interact with them and become aware of their value added. Based on the previous feedback, this was especially applicable for filters. We added a quick <strong>“what’s new” message</strong> that pops up during the first time that a user visits the page.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*ixjpyyL6IOKbKn4D7wrQHA.png" /><figcaption>If your app’s release notes are not integrated into your platform yet, “what’s new” messages are a great way to communicate big changes.</figcaption></figure><p>By monitoring the first few weeks of usage, we found that users were divided into two groups:</p><ul><li>One was using purely the search input without filtering their list.</li><li>The other had a more balanced usage in both search input and filters applied.</li></ul><p>The investigation led us to the reason: it <strong>had to do with the demos we conducted.</strong> All users from the first group did not participate in any demo. Users from the second group took part of at least one of them. This meant that the “what’s new” message was not enough to change user behavior.</p><p>To improve awareness, we targeted new places in the UI where we could inform users about it. A good place to start was the <strong>“no results found” message.</strong> If that group of users were performing so many searches, they were landing on that page often. We substituted “Try reducing your search term and apply less filters to get better results” to what you can see in the screenshot below:</p><figure><img alt="No results message" src="https://cdn-images-1.medium.com/max/1024/1*Csn9PlWpnvO_2EunrosWfw.png" /><figcaption>Use the “no results” message to help your users to recover from their possible errors</figcaption></figure><p>Within 3 months, we’ve launched our new product search, available for all our users. This release is a great symptom of the maturity that the Merchant Center is approaching. And that would not be possible without our users’ help. Thanks to the input and user’s feedback we collected throughout the process, we are confident the new search will improve the search experience for every catalog manager.</p><p>______</p><p>Thanks to our testers, beta users, designers, developers, product owners and anybody who helped building this feature. And special thanks to Andrea, Brad, Freddie, Adnan and Yann for their contributions to this article.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=7f8ac75dbeba" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[How to reduce the distance when the UX designer works remotely]]></title>
            <link>https://medium.com/@mbarrena/how-to-reduce-the-distance-when-the-ux-designer-works-remotely-ffdc6afa77f?source=rss-57049ad8ac6b------2</link>
            <guid isPermaLink="false">https://medium.com/p/ffdc6afa77f</guid>
            <category><![CDATA[remote-working]]></category>
            <category><![CDATA[user-interface]]></category>
            <category><![CDATA[ux]]></category>
            <category><![CDATA[user-experience]]></category>
            <category><![CDATA[remote-work]]></category>
            <dc:creator><![CDATA[María Barrena]]></dc:creator>
            <pubDate>Tue, 05 Mar 2019 12:58:50 GMT</pubDate>
            <atom:updated>2023-04-01T17:53:50.644Z</atom:updated>
            <content:encoded><![CDATA[<h3>Reducing the distance when the UX designer works remotely</h3><h4>Ideas and techniques based on my experience as a UX/UI designer, working remotely with my vertical team.</h4><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*itpAdE6O7ldkLqBqvW2ZTA.jpeg" /></figure><p>As my colleague Ida <a href="https://techblog.commercetools.com/vertical-teams-and-collective-thinking-4d4943112588">explained,</a> my product team embraced vertical, cross-functional teams. Each team has its own UX/UI designer that focuses on the team’s area of responsibility. I am one of them. However, we are not sitting in the same office as the rest of our team members.</p><p>Even though the situation was initially inconvenient, our team has overcome this difficulty. These are the steps we have followed to succeed on a better remote collaboration, with special focus on my role as UX designer.</p><h3><strong>Start in person. Continue remote</strong></h3><p>Small teams create meaningful connections and build trust. But how do we establish these connections while apart?</p><p>Building good human relations is harder when you are not in the same office. That’s why three weeks after I joined the team, I decided to visit them. Starting the relationship in person sets up the solid foundation of your professional collaboration. Having the team together in one place as frequent as possible has brought us a good outcome.</p><blockquote>Starting the relationship in person sets up the solid foundation of your professional collaboration.</blockquote><p>Once you are in the same room, there are many activities you can do to reduce the technical and personal gap. Here are the ones that I’ve tried out and have helped:</p><h4><strong>1. Get to know each other better</strong></h4><p>The first point is quite obvious and you should apply it either you are in the same room or not. You need to get to know your team better and vice versa. Organise an activity all together outside of work, in a more relaxed environment. Knowing somebody more personally always creates empathy. It is also a great opportunity to refute wrong myths they might think about the role of a UX designer. Ask about their previous job experience and their experience working with designers.</p><h4>2. Bring UX mentality to the team</h4><p>Even though the role of UX and UI designers are not new in the IT industry, we need to explain often the differences between those two to other stakeholders nowadays. Some people still believe that design is only about making a feature nice and happens at the end of the chain. This type of thinking is often associated with waterfall models of software development. It can get worse if all the UX/UI designers of a company are sitting together far from the rest. Teams can end up isolating that role as they will see the design department as an external agency. The sooner you start working towards changing such mentality, the better.</p><p>Share with the vertical team any UX related task you would like to include into your design iterations and the team workflow. Explain what is it for and the benefits for each team member this can bring. It can be a certain way of doing research, a specific type of test or any other tactic. Giving examples of successful use cases can help you to convince the most skeptical ones. Showing the positive aspects of your area of expertise proves that you are there to help as well. If that works out, share the results of that task with the team afterwards.</p><p>During my first few months working with the team, I’ve suggested conducting usability tests. The goal was to verify whether the proposal we were discussing and working on was the right approach. As none of us were the real end user, it was a good task to perform. Once I performed those with real users, I’ve presented the learnings to the product team. This guarantees that the design decisions taken are user approved before implementation. It can also inspire other teams to follow similar design strategies.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*4nm1FUX80KuhWfj6rPh99g.jpeg" /><figcaption>Brainstorming can show you a completely different perspective and solution to the same problem.</figcaption></figure><h4><strong>3. Brainstorm ideas with all the team</strong></h4><p>Get a whiteboard and draw your proposals for the feature that you are working on. Then discuss it and brainstorm with the entire team<strong>.</strong> Inviting everybody to “mess-up” with your sketch shows a willing to listen and trust on their feedback. Don’t forget to explain what consequences their suggestions will bring to the user. It’s a great way to start spreading UX culture within the team.</p><p>Although a bit more challenging, you can try this exercise remotely as well. But as it happens in usability tests, being able to see everybody’s physical reactions is as important as what they say. That’s why it is beneficial that the first time you try this out within the team, you do it in person.</p><h4><strong>4. Don’t be afraid of code</strong></h4><p>Since I started working in a web environment, I have read the same quote over and over again. “If you are a designer, you need to know how to code”. I believe a more accurate rule that UX and UI designers should apply is the following:</p><blockquote>If you are a designer, you need to understand each stakeholder’s complexities. That’s our job in a nutshell.</blockquote><p>There are many ways to get to know the technical limitations developers face. An effective one is sitting down with a developer and to observe how they work and which tools they use. This technique is called “shadowing” and it is used to train people in other industries. When new experts start working in an Apple store, they have to spend a week shadowing a senior specialist. That’s the way for them to learn more about the do’s and don’t of the job.</p><p>If you are a developer reading this, bring the designer to technical meetings. The sooner you do it, the earlier they will start to become familiar with technical concepts. Terms that they will see in technical documentation when they start digging more on it. It can help designers to understand the complexities a design proposal can bring.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*h-1dafU3Mcq54UA_CP2WVQ.jpeg" /><figcaption>Taking five minutes to explain a technical detail can save both the developer and the designer time afterwards.</figcaption></figure><p>UX designers will identify all these techniques as some of the best practices in user testing. What nobody in our industry dares to say is that we have more than one user to think about. It is clear that the tasks that we work on on a daily basis have the end user in mind. But our own team-mates are another type of user as well: the users of our design proposals. It’s to them, after all, that we present new ideas to and receive our deliverables from. A UX designer needs to understand the vertical team motivations and goals. You should never forget about all those when working on a task.</p><h3><strong>(Bidirectional) Communication, communication and communication</strong></h3><p>Once you are back to your daily routine, the challenge is to make that bond that you have established strong. Being far away is not an excuse for the team to forget about who’s on the other side or not keeping a daily contact with them.</p><p>At first, you might need to establish the habit by reminding you are still there. Ask to take part in the team meetings, even if they might sound technical at first. In those, you can learn the common shared language. Avoid that dev-design gap that happens often as soon as possible. Expressing design decisions using the same vocabulary will avoid confusion. Don’t be afraid to ask what you don’t understand. Most of the people are happy to explain to you what they know and do.</p><p>After a few weeks, the habit of having a continuous contact will be established. The challenge then consists on being consistent with the bidirectional contact.</p><p>Our vertical team follow a Scrum way to work without estimations. This type of structure has moved us away from a more sequential development. We do two-week sprints with sprint plannings, daily standups (with a physical board), retrospectives and refinment sessions.</p><blockquote>Refinement sessions are one of the most essential meetings we hold.</blockquote><p>Refinement sessions are one of the most essential meetings we hold. We go through the upcoming frontend tasks inside our backlog and refine them. We review the user flow and interaction we have already agreed upon and take a look at the proposed UI. Backend developers are responsible of checking the API implementation. They can confirm whether the backend allows us to build that or should be changed. Then, we write down the acceptance criteria with the help of the product owner. We also verify that the task does not need to be split into smaller tasks.</p><p>In most cases, this is the first time that all team members are aware of what’s coming up. Everybody takes a more active role by discussing how something needs to be built.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*PLZWn9zhOobxum16DPiynA.jpeg" /><figcaption>Continuous communication, the most essential part of remote collaboration</figcaption></figure><p>We also have separate meetings when we need to discuss a certain feature or task. We go over the design proposals with at least one frontend and one backend developers and the product owner. In those, we take meeting notes that we store so we can always go back to them.</p><p>This type of structure helps us having a continuous communication within the team. But it can be complex if the technical setup is not a good one. The equipment ought to provide good video and sound quality. Avoid wasting time with technical issues as much as possible.</p><p>As with any relationship, the effort of not losing the contact should come from both sides. Leaving somebody apart, even the far one does nothing but harm. Nobody should forget that a designer can help with their expertise on any of the steps of the project. By not conferring with the designer, you are communicating to the team UX is not relevant at all. Furthermore, you are giving the designer the impression that you don’t trust in their expertise.</p><h3><strong>Writing more is not a waste of time</strong></h3><p>When you rely on internet connection, bringing the entire team to a meeting can be hard. Once of the skills that I’ve been improving a lot since I work on remote is writing proposals.</p><p>When the conference call fails, the design decisions in written form is my backup plan. In those writings I provide a detailed explanation. They include the pros and cons for the user and describe the edge cases that I can oversee. Of course, those documents include pictures of sketches, screenshots and/or digital prototypes. Afterwards, I’ve shared that with the team.</p><p>This has been a good exercise for me to get to synthesise ideas, being as clear as possible. This is quite essential when you need to present proposals on a conference call. Discuss something in written form is like having meeting notes without the meeting.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*HXI010qfM4LhiYRURS-fTQ.jpeg" /><figcaption>Paper prototypes are harder to use to communicate ideas remotely</figcaption></figure><h3>Getting feedback from far away</h3><p>Giving and receiving rapid UX feedback being remote is not as easy as in person. The fact that I’m not in the same room has made me jump from paper prototypes to digital ones. I use digital prototypes during meetings to show the possible user flow and layout.</p><p>Most of the times, the round of feedback happens during the meetings scheduled earlier on. This makes the design iterations longer than normal. But the biggest disadvantage in this case is the lack of useful feedback that happens often.</p><blockquote>Examples of unhelpful feedback sentences are:<br>“We cannot build that”,<br>“The answer is too technical for you”<br>“It is easier to explain if I could go to your desk”</blockquote><p>No matter your role, always prioritise giving constructive feedback to your colleagues. No excuses. Writing a short message on a chat or making a quick video call to explain something takes less than 5 minutes.</p><p>Even though the reasons might be too technical, get interested in those. It is your responsibility to learn more about the limitations the team faces. Your design iterations will get shorter. This becomes essential for the teams that have to deal with deadlines.</p><h3><strong>Results after one year</strong></h3><p>More than one year has passed since we are working with such setup. I must say, I’m proud of the level of collaborative thinking the team have achieved so far. Nobody works <em>solo </em>and we always bring expertise to the table. I’m quite confident that my work wouldn’t differ much if I were sitting in the same room with my teammates. We keep gathering together from time to time to not lose the human touch. I’ve visited them once every few months while some of my teammates have visited me at least once.</p><p>Together as a team we have reached a consensus of how we want to work. We find solutions to the problems together as a team. We have built trust between each other. Trust in providing feedback or discussing points of friction. And above all, we take care of not leaving the remote team member behind, no matter what their role in the team is.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=ffdc6afa77f" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[UX research for a music festival app(a Capstone Project and an Interaction Design Course)]]></title>
            <link>https://medium.com/@mbarrena/a-music-festival-app-a-capstone-project-and-an-interaction-design-course-11a6f8762a91?source=rss-57049ad8ac6b------2</link>
            <guid isPermaLink="false">https://medium.com/p/11a6f8762a91</guid>
            <category><![CDATA[design]]></category>
            <category><![CDATA[interaction-design]]></category>
            <category><![CDATA[ux]]></category>
            <category><![CDATA[music]]></category>
            <category><![CDATA[coursera]]></category>
            <dc:creator><![CDATA[María Barrena]]></dc:creator>
            <pubDate>Wed, 03 Aug 2016 19:27:52 GMT</pubDate>
            <atom:updated>2020-10-19T10:16:14.463Z</atom:updated>
            <content:encoded><![CDATA[<p>As a music festival lover living right in the center of Europe, I have been always open to attend to music festivals on any place that I can reach easily by bus, train or even plane. That means for me any European country. I’ve attended to music festivals in Germany, Austria, Spain, Portugal and Sweden already. But unfortunately it is quite hard to know, find and track all music festivals happening in such a big extension. It takes time and effort and not everybody is interested in spending their time doing that research. Right now, in 2016, there is no app and no updated website that offers that type of information.</p><p>Professionally, I recently switched jobs from graphic designer to interaction designer. My goal is to get better and better in the UX field. That’s why I’ve been studying this <a href="https://www.coursera.org/specializations/interaction-design">Interaction Design Specialisation</a> that Coursera added to their platform last year and gave me the chance to make that change. When I learnt I would be making a final project for the course, I was excited to develop the idea a little further as this one has been in my head for quite some time already. I didn’t know by then, but this is what can be called a “user need”.</p><p>I decided to verify if that was a problem for other people too. By conducting interviews I found out that despite the answer was yes, most of the people find the bands attending secondary (the don’t have to really love them). Being with their friends was the primary reason why they would attend to a specific music festival. That put me in the direction of thinking that the social component should be as important or even more as the bands playing.</p><p>The first step was making initial storyboards with possible scenarios like the one described and problems that people might have. Paper prototypes with screens solving those problems came afterwards. They are extremely useful when you are trying to decide how the user flow should be.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/740/1*UlF0IEFUoliWWG426D1Djw.png" /><figcaption>First paper prototypes</figcaption></figure><p>I translated those to digital mockups. I started developing two different versions in parallel of a possible app. The user would be able to track artists and friends so the app can suggest you which music festivals you might be interested in, but one of the versions would focus more on the music festivals your friends are interested in going, rather than making you suggestions based on your music taste.</p><p>Thanks to the Heuristic Valuation given by a fellow UX designer, I discovered that the onboarding area of the app should be quite relevant as that’s the entry point where the user would like to add the bands they would like to see live and connect the account with any social media account that the user uses. That area should be quite straight forward. At this point, both versions were combined into one and I had to dismiss some ideas to keep it as simple as posible. Product managers tend to talk about a minimum viable product (MVP) and as I wouldn’t be able to code anything, I had to stick with that idea.</p><h4>The Digital Prototype</h4><p>After thinking on the whole app structure, it was time to start building up the digital prototype. The goal was to test it with people. I decided to use Axure for that, as it’s a program that I am already using it for work. It would take me less time to build if I already know the tool. By using Axure and the Wizard-of-Oz technique where the prototype simulates it’s working but it’s actually not, I was ready for the next step: testing the idea with real users.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*t2oEM3gtVWS_nfUCHrY7IA.jpeg" /><figcaption>First test with targeted users</figcaption></figure><p>I had the chance to test it with two types of users: one that finds the festival where the group will go and one that would go where all his friends are planning to attend no matter what bands are playing. The feedback was quite good as it made me learn one of the most relevant thing of the app, the option to follow bands, wasn’t self explicative and needed a different copy.</p><p>Another piece of feedback that I got was related to the social part of the app. I thought it would be nice to have a chat option for all users attending to each specific festival. It can be used to share info, sell tickets, share a car or simply talk about the bands playing. This button was somehow misleading, as the users who tested the app thought that chat was among their friends only, rather than any person attending to the festival.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*z0zYuzLdHsW7k0cMStViMw.png" /><figcaption>Performing one of the A/B tests</figcaption></figure><p>In order to verify if that was the same thought for other people, I conducted an A/B test with a different approach for the chat, adding a better copy to explain what that chat was about. My initial bet was there wasn’t a statistically significant difference between A and B and my hypothesis was I would get zero clicks in the chat button on both cases.</p><p>The results verified my guessing. The users didn’t really have a problem with that area of the app but with the onboarding, as they didn’t expect something like that once they have started the app for the first time.</p><p>After fixing a couple of other details, the prototype was done and ready. You can get an idea about the app by checking this video that I made. It includes the user need, and how the app helps people that enjoys music festivals as much as me.</p><p><a href="https://www.youtube.com/watch?v=mDQKhy2wR2Y">https://www.youtube.com/watch?v=mDQKhy2wR2Y</a></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=11a6f8762a91" width="1" height="1" alt="">]]></content:encoded>
        </item>
    </channel>
</rss>