<?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 Josh Winters on Medium]]></title>
        <description><![CDATA[Stories by Josh Winters on Medium]]></description>
        <link>https://medium.com/@subvocalizing?source=rss-2971933d869e------2</link>
        <image>
            <url>https://cdn-images-1.medium.com/fit/c/150/150/1*27VWxb7BpclskteTbNBMWA.jpeg</url>
            <title>Stories by Josh Winters on Medium</title>
            <link>https://medium.com/@subvocalizing?source=rss-2971933d869e------2</link>
        </image>
        <generator>Medium</generator>
        <lastBuildDate>Sat, 21 Nov 2026 14:03:48 GMT</lastBuildDate>
        <atom:link href="https://medium.com/@subvocalizing/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[Principles for Picking the Perfect Logo Font]]></title>
            <link>https://medium.com/@subvocalizing/principles-for-picking-the-perfect-logo-font-b3f47da2d441?source=rss-2971933d869e------2</link>
            <guid isPermaLink="false">https://medium.com/p/b3f47da2d441</guid>
            <category><![CDATA[chatgpt]]></category>
            <category><![CDATA[design]]></category>
            <category><![CDATA[logo]]></category>
            <category><![CDATA[logo-design]]></category>
            <dc:creator><![CDATA[Josh Winters]]></dc:creator>
            <pubDate>Tue, 27 Dec 2022 22:25:18 GMT</pubDate>
            <atom:updated>2022-12-27T22:27:20.917Z</atom:updated>
            <content:encoded><![CDATA[<h3>A Guide to Choosing the Ideal Logo Font for Your Brand</h3><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*NbkhG3BbT_VMVAY6" /><figcaption>Photo by <a href="https://unsplash.com/@emilybernal?utm_source=medium&amp;utm_medium=referral">Emily Bernal</a> on <a href="https://unsplash.com/?utm_source=medium&amp;utm_medium=referral">Unsplash</a></figcaption></figure><p>When picking a logo font, it is important to consider factors such as legibility, size, style, scalability, compatibility, distinctiveness, cost, trends, relevance, and versatility. Selecting the perfect font for your brand requires careful thought and consideration. Make sure the font you choose reflects the values and message of your brand identity, is legible and distinct, and is appropriate for multiple applications. With careful consideration and research, you can find the perfect logo font to represent your brand.</p><h4>1. Legibility: Make sure the font is easy to read and understand.</h4><p>Legibility is key when it comes to selecting a logo font. After all, if a font is hard to read, it will be ineffective for conveying your brand’s message. Before settling on a font, take the time to evaluate its legibility, both onscreen and in print.</p><h4>2. Size: Choose a font size that is large enough to be readable even when scaled down.</h4><p>Make sure to choose a font size that is large enough to be readable even when scaled down. This will ensure that your logo font is visible and recognizable at any size. Consider the smallest size your logo might be used at and make sure the font is still legible.</p><h4>3. Style: Select a font that fits with the style of your brand identity.</h4><p>When selecting a font for your logo, it is important to consider the style of your brand identity. Choose a font that will work with the overall look and feel of your brand’s aesthetic. If your brand is modern and modern, consider a contemporary font. If your brand is more traditional, pick a font with classic typefaces.</p><h4>4. Scalability: The font should be able to be scaled to different sizes without losing its legibility.</h4><p>Scalability is also an important factor when it comes to logo fonts. The font you choose should be able to be scaled to different sizes without losing its legibility. It should also be compatible with the design software you are using.</p><h4>5. Compatibility: Make sure the font is compatible with the design software you are using.</h4><p>Make sure the font you choose is compatible with the design software you are using. This will ensure that your logo looks the same across all platforms. Consider the various software you will be using and select a font that is compatible with each.</p><h4>6. Distinctiveness: Choose a font that stands out and is unique enough for your logo.</h4><p>It is important to choose a font that stands out and is unique enough for your logo. Make sure the font you select is not too common and easily recognizable. Avoid fonts that are too similar to popular fonts like Helvetica or Arial.</p><h4>7. Cost: Evaluate the cost of the font before making a final decision.</h4><p>Before making a final decision, evaluate the cost of the font. Make sure you are getting the best value for your money. Consider free fonts as well as fonts that require a license.</p><h4>8. Trends: Select a font that is timeless and doesn’t follow any current trends.</h4><p>Select a font that is timeless and doesn’t follow any current trends. This will help your logo stand out and remain relevant for a longer period of time. Avoid overly trendy fonts that could quickly become outdated.</p><h4>9. Relevance: Make sure the font is relevant to the type of brand you are creating.</h4><p>Make sure the font you choose is relevant to the type of brand you are creating. The font should reflect the values and message of your brand identity. Consider the tone and style of your brand and select a font that fits the overall theme.</p><h4>10. Versatility: Choose a font that can be used in multiple contexts and applications.</h4><p>Lastly, choose a font that is versatile and can be used in multiple contexts and applications. This will help to ensure that your logo looks great in any situation. Consider how the font will look on print materials, digital platforms, and other media.</p><p>Choosing the right font for your logo is an important decision. Make sure to consider all the relevant factors and evaluate your options carefully. With the right font, your logo will look great on any platform, be recognizable and distinct, and be an accurate reflection of your brand identity. With these principles in mind, you can find the perfect font for your logo.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=b3f47da2d441" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Questions to ask founders when joining a seed-stage startup]]></title>
            <link>https://medium.com/@subvocalizing/questions-to-ask-founders-when-joining-a-seed-stage-startup-4ef0ebd63bdf?source=rss-2971933d869e------2</link>
            <guid isPermaLink="false">https://medium.com/p/4ef0ebd63bdf</guid>
            <category><![CDATA[startup]]></category>
            <category><![CDATA[business]]></category>
            <category><![CDATA[interviewing]]></category>
            <category><![CDATA[equity]]></category>
            <category><![CDATA[founders]]></category>
            <dc:creator><![CDATA[Josh Winters]]></dc:creator>
            <pubDate>Sat, 15 Feb 2020 01:34:44 GMT</pubDate>
            <atom:updated>2020-02-15T01:38:36.968Z</atom:updated>
            <content:encoded><![CDATA[<h4>If you are dedicated to building a business from scratch as an early employee, know what you are getting into before joining.</h4><p>After four years trying to get a big idea off the ground at a seed-funded startup, I am back on the interview circuit. When I was interviewing with my previous employer, I had very little context on how to evaluate the founders or business. Since then I have learned a tremendous amount about the management of startups, and I would like to begin sharing my knowledge with the startup community that has been so beneficial to me throughout my journey.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*-Y9ozbNWSViiCmal1TT32w.jpeg" /><figcaption>Photo by <a href="https://unsplash.com/@dsmacinnes?utm_source=unsplash&amp;utm_medium=referral&amp;utm_content=creditCopyText">Danielle MacInnes</a> on <a href="https://unsplash.com/s/photos/startup?utm_source=unsplash&amp;utm_medium=referral&amp;utm_content=creditCopyText">Unsplash</a></figcaption></figure><p>My first post in this vein, and most relevant to those interviewing, is a list of questions focused on the evaluation of seed-stage startups. They will help you gauge the maturity of the business. By no means is this an exhaustive list and you might not need answers to all of them prior to joining. All of these questions are entirely reasonable to ask a founder, especially in a pre-product market fit startup, and the founder should provide straightforward and honest answers. Some founders might want an NDA in place prior to answering questions about revenue or customers (which is fine but read carefully, should be scoped around a potential business relationship), and some customers prefer their names not be shared (in which case a general description of the customer should suffice, but be sure you are discussing real, paying customers).</p><p>The goal is not to eradicate risk, but instead to make wise decisions about where to invest your time and energy. Consider any misrepresentation you discover after joining a red flag and explore better opportunities.</p><p>If you have suggestions for additional questions, or would like clarification on anything, please respond in the comments.</p><p><strong>Business</strong></p><ul><li>Can you (founders) tell me about your background in your industry?</li><li>Why did you start this business?</li><li>How did you connect with your investors and what inspired them to invest in you? (Pay close attention to investor quality — look for prior exits in/around the startup’s space)</li><li>What are your revenue streams?</li><li>What is your pricing model?</li><li>What is your average MRR over the last three months?</li><li>What are your average monthly costs/expenses? COGS, burn</li></ul><p><strong>Customers</strong></p><ul><li>Who is your target customer, and why? (For pre-PMF, you want an answer that sounds <em>too </em>precise)</li><li>How many customers do you have?</li><li>How long have you had them?</li><li>What are they paying for?</li><li>How much are they paying?</li><li>Who is your oldest customer?</li><li>How are you acquiring customers?</li><li>How are they responding to your product?</li><li>How do you research your customers?</li></ul><p><strong>Product</strong></p><ul><li>How much of what you have on your marketing site has been built? What has yet to be done?</li><li>Can you show me your product?</li><li>Who are your primary competitors?</li><li>Who is the leader in your space?</li><li>Who are the biggest players in your space?</li><li>Are any specialities standing out for you?</li><li>What is your growth strategy and how might your strategy fail?</li><li>What parts of your product are closest and furthest from the ideal in your vision?</li></ul><p><strong>Team</strong></p><ul><li>How has your team changed over time?</li><li>Which employees have been around longest?</li><li>Can you describe your working relationship with your employees?</li><li>(If remote) What tools and processes do you use to stay connected with remote employees?</li></ul><p><strong>Code </strong>(questions for tech lead)</p><ul><li>Have you been tracking code quality?</li><li>What development tools do you use?</li><li>What version control do you use?</li><li>Can you describe your stack for me?</li><li>How is your software organized?</li><li>How do you approach testing?</li><li>What is your deployment process?</li></ul><p><strong>Equity</strong> (when discussing compensation)</p><p>Equity is complicated and you should try to understand your equity compensation as fully as possible. Do your research and understand what you are getting. Other resources explain much more thoroughly than I could, but at minimum you should understand the percent of equity your options represent, current strike price, vesting schedule and whether everyone else is on the same schedule, and liquidation preference.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=4ef0ebd63bdf" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Advice for Engineers]]></title>
            <link>https://medium.com/@subvocalizing/advice-for-engineers-81c5ec42b37f?source=rss-2971933d869e------2</link>
            <guid isPermaLink="false">https://medium.com/p/81c5ec42b37f</guid>
            <category><![CDATA[programming]]></category>
            <category><![CDATA[software-craftsmanship]]></category>
            <category><![CDATA[startup]]></category>
            <category><![CDATA[software-development]]></category>
            <category><![CDATA[software]]></category>
            <dc:creator><![CDATA[Josh Winters]]></dc:creator>
            <pubDate>Fri, 27 Dec 2019 20:05:20 GMT</pubDate>
            <atom:updated>2020-01-17T00:02:05.968Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*j5zmGtsPMfHSjrxEQOArQw.png" /><figcaption>Work at the boundary of noise and order</figcaption></figure><p>Stream of thought — bear with me!</p><p>Don’t be clever. Keep the system as simple as possible. Push complexity into the leaves of the codebase where it can be contained and managed until a solid pattern emerges. When starting on a new system, make it work end-to-end before trying to making it elegant or filling in auxiliary methods. Always know how to back out of a change. Plan architecture around flex points. Plan for change in both feature-level functionality and core product value. Be one step ahead of the product team in generality, but no more, so that you can change course quickly when the business knows which way to head. Don’t assume you know where the business is headed (beyond a reasonable horizon).</p><p>Look for passion in yourself and work where you are passionate. Find a niche that fascinates you. Right industry, right company, right role, right product, right layer, right module, right function. Invest in your career like like a value investor picks stocks. When you have thoroughly considered your position and know you are in the right place working on the right thing, work is blissful (as much as it can be). It takes time and effort to find your best position, and it changes as you grow.</p><p>Procrastination has a meaning — are you taking care of yourself? Your workspace?</p><p>When you do something awesome, take some time to enjoy and reflect on it before moving on — what you liked about it, what you would do differently if you started again, where it fits into your broader scheme.</p><p>Code automates work, so think in terms of how much work you are automating. Develop an accurate picture of how much work you are automating — it is easy to think you’re helping when you are way off track. Usually there is a way to organize work so that you’re solving multiple problems at once — this is the art of engineering, and you only recognize these possibilities with practice and mindfulness.</p><p>Data has potential value, but only when organized and stored properly and consistently. Value must be carefully extracted from potential.</p><p>Engineering is not the only business discipline that requires planning, construction and maintenance.</p><p>For greatest learning potential, own whatever you are working on as fully as possible — but beware, this is extremely stressful. There is no right answer in life, everyone is making up stories as they go along. If you don’t like the path you are on, you have three options: do nothing, course correct, or find another path altogether.</p><p>Your creative effort is worth money unless it’s for personal gratification, in which case money is optional (but usually desirable and sometimes attainable). Money isn’t everything, but if you’re underpaid too long you will grow resentful. There are a lot of lazy people in the world who want you to do all their work for them, and they will be remorseless in extracting life force from you if you let them. Charge in proportion to your expected effort and discount on desire to do the work.</p><p>When you’re interviewing to join a business full time, ask hard questions and expect the right answers. Be respectful but have standards. Do your due diligence and bear in mind that some people are good at hiding poor or negligent performance. When in a position to hire, do so carefully and methodically.</p><p>Take your lumps and take risks. Well-considered risks pay off surprisingly often. Accept the potential for failure, which comes in two flavors — causally attributable and arbitrary bad luck. When you’re wrong, take responsibility. Don’t repeat your mistakes.</p><p>Everyone wants a friend and even in difficult cases you can be a good friend. The people around you are doing their best, whether they are succeeding or not.</p><p>Family comes first.</p><p>Exercise and fresh air are intrinsically rewarding.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=81c5ec42b37f" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Making the Most out of Vehicle Traffic Counts]]></title>
            <link>https://medium.com/idealspot/making-the-most-out-of-vehicle-traffic-counts-30349907dda7?source=rss-2971933d869e------2</link>
            <guid isPermaLink="false">https://medium.com/p/30349907dda7</guid>
            <category><![CDATA[vehicle-tracking]]></category>
            <category><![CDATA[vehicle-traffic]]></category>
            <category><![CDATA[transportation]]></category>
            <category><![CDATA[brick-and-mortar]]></category>
            <category><![CDATA[retail]]></category>
            <dc:creator><![CDATA[Josh Winters]]></dc:creator>
            <pubDate>Wed, 22 Nov 2017 00:31:10 GMT</pubDate>
            <atom:updated>2017-11-22T00:33:52.894Z</atom:updated>
            <content:encoded><![CDATA[<p>6 steps to evaluating a location’s vehicle traffic potential.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/300/0*H9BlRzb7L0uqb0H_.png" /></figure><p>Vehicle traffic is key to a successful retail location. However, not all traffic is created equal. Commuters, lunch crowds, and weekend shoppers all consume products and services in different ways. You need traffic data that is accurate and detailed enough to identify different groups of customers so you can make decisions accordingly.</p><p>We have been helping businesses find and understand traffic counts for years, and we’re excited to share the methods we’ve discovered to getting the most value out of your traffic counts.</p><h3>1. Know your average daily traffic</h3><p>The industry standard <strong>Average Annual Daily Traffic (AADT)</strong> shows the typical vehicle traffic per day in front of your location.</p><p>Knowing AADT is like knowing the number of residents in your market. You must be aware of this number to understand your maximum customer reach. If 100% of the customers who drove past your business stopped and spent $10, how much money would you make per day? Would it cover your daily costs? If you can’t cover your costs with every single vehicle stopping and spending money (which is highly unlikley to happen), you probably shouldn’t locate along that road.</p><p>Of course, not everyone will stop. You’ll want to know the percentage of total residents in your target market that fit your expected demographic and are interested in purchasing what you have to sell. In most cases, you can multiply this percentage by your average daily traffic and get a rough estimate of how many vehicles will stop to visit you per day. (Highways and large arterial roads should be excluded from this method since they carry through-traffic from further away).</p><h3>2. Reach your customers at the right time of day</h3><p>Most experts will tell you that AADT is only the beginning of the story.</p><p>Let’s consider a scenario. Say you’re opening a coffee shop for morning commuters and are looking for a heavily trafficked street to locate on. You look around and finally find a strong prospect with 12,000 vehicles per day. It’s priced right and seems to be in good condition! Great location, right?</p><p>Well, maybe. Those 12k vehicles could be evenly spread throughout the day, cutting the morning traffic down to three thousand. Still not bad, but thorough consideration of traffic at other locations may show that while another location has a lower AADT, <em>morning</em> traffic counts are <em>higher</em> than three thousand elsewhere. Furthermore, since the average daily count is lower at our hypothetical comparable location, rent may be lower as well because marketers can’t use it as a selling point. Without time of day counts, you might miss out on a great deal that would have netted you more customers because you didn’t fully understand your location’s morning traffic. And of course, this example can apply to other times of day as well. That’s why we recommend <strong>time-of-day vehicle traffic counts</strong> in addition to AADT.</p><h3>3. Be on the best side of the street</h3><p>Let’s continue with the coffee shop example. You found a location with high morning traffic counts. Great! But we’re not done yet; you need to know <strong>side of street traffic</strong> before you can nail this one down. Morning commuters generally head in one direction: toward work. Often this is toward an urban area, away from residential areas. If you’re choose the wrong side of the street, you’re erecting a huge obstacle between yourself and your customers. For your customers to reach you, they must:</p><ol><li>See your sign from across a busy road</li><li>Navigate into the turn lane</li><li>Wait until they can cross the street. This may involve going through a light (potentially waiting through several light changes), oncoming traffic, pedestrians, or otherwise backing up the traffic behind them.</li><li>Cross the street, sometimes requiring a U-turn that causes them to go backwards — AWAY from their final destination.</li><li>Pull into your lot, buy your delicious coffee, leave.</li><li>Turn out of your lot and make their way back to the other lane (with all the same barriers).</li><li>Finally continue on their path to work.</li></ol><p>We’ll tell you right now: that’s a lot of effort for a $3 cup of coffee! Instead of causing that collective hassle and inefficiency, you should instead aim to be on the same side of the street as your customers from day one.</p><p>Last thought here — if you’re a coffee shop on the wrong side of the street, and a competitor moves in on the right side of the street, you’re in for a difficult fight. Nip that one in the bud if you can.</p><h3>4. Compare weekdays to weekends</h3><p>AADT only provides an average between weekdays and weekends. Again as a coffee shop, you’ve decided your target audience will be morning commuters. Do you think they’ll be driving by on weekend mornings, when they’d rather be sleeping in, eating brunch at a local diner, and spending time with their families? You’re going to spend a lot of time, energy, and money developing your brand as a stop for morning commuters. This is a competitive segment and if you’re on a road better suited for weekend warriors, then you’re not reaching the audience for whom you are spending valuable ad dollars. <strong>Day of week traffic counts</strong> will help you distinguish between these segments.</p><h3>5. Use high-quality, recent data</h3><p>Most public counts track average annual daily traffic, but not side of street, day of week, or time of day traffic counts. However, with the advent of mobile location tracking technology stronger traffic data sets are emerging across the industry. For instance, our provider, <a href="http://inrix.com/">INRIX</a>, tabulates their counts from 300 million devices and vehicles across 1.8 million roads throughout the United States.</p><p>If you’re not careful, you can accidentally make decisions based on severely outdated information. Most public and private traffic data providers include counts up to <em>a decade old</em>. Let us remind you that a decade ago the iPhone first launched, the housing bubble was approaching its zenith, and George W. Bush was still president!<em> </em>INRIX, on the other hand, only uses data from within the last three years, and metropolitan areas tend to have much more recent counts because more devices and vehicles are being tracked every day.</p><p>One final point to consider when assessing data is that the methodology for counting traffic has changed significantly over the last few decades. Historically, counts would be taken by someone standing on a corner with a clicked for eight hours. Over time that transformed into a rubber tube laid on a road for a couple days that increments a digital count when a set of tires rolls over it. Now with device and vehicle location tracking, we have 24/7 digital counts every week of the year, often along with origin and destination statistics. Which method would you trust most?</p><h3>6. Compare several locations</h3><p>As we mentioned above, you might come across a location that looks like it has a lot of traffic due to high average daily counts, but which won’t put you in front of the right consumers depending on time of day, side of street, or day of week. There are also different <em>kinds </em>of roads that urban planners use when drawing up plans for cities. Highway traffic is very different from traffic on a local residential road. You might find that being a community staple on the only road out of a neighborhood suits you better than sitting in a large shopping center off a busy highway.</p><h3>IdealSpot Vehicle Traffic Reports</h3><p>That’s a lot of information to consider. Not only do you have to find clean, high-quality, up-to-date data, but you have to find counts for every time of day, day of week, and side of street for any location you’re interested in!</p><p>Fortunately, we at IdealSpot have already solved these problems for you. Our vehicle traffic reports include Average Annual Daily Traffic (AADT), Day of Week, Time of Day, and Side of Street traffic using the most up-to-date INRIX traffic counts. You can also compare up to 10 other roads in one report.</p><p>The data we provide normally costs large brands hundreds of thousands of dollars per year to access, let alone visualize and digest. We consider $300 per report a bargain and guarantee our reports are well worth your investment! <a href="https://www.idealspot.com/traffic/">Find traffic counts in less than five minutes through our self-service report builder!</a></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=30349907dda7" width="1" height="1" alt=""><hr><p><a href="https://medium.com/idealspot/making-the-most-out-of-vehicle-traffic-counts-30349907dda7">Making the Most out of Vehicle Traffic Counts</a> was originally published in <a href="https://medium.com/idealspot">IdealSpot</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Comparing Spending and Demand in Site Selection]]></title>
            <link>https://medium.com/idealspot/at-idealspot-we-speak-a-lot-about-how-important-understanding-local-demand-is-for-your-business-c91c147b12d0?source=rss-2971933d869e------2</link>
            <guid isPermaLink="false">https://medium.com/p/c91c147b12d0</guid>
            <category><![CDATA[retail]]></category>
            <category><![CDATA[demand]]></category>
            <category><![CDATA[site-selection]]></category>
            <category><![CDATA[consumer-behavior]]></category>
            <category><![CDATA[idealspot]]></category>
            <dc:creator><![CDATA[Josh Winters]]></dc:creator>
            <pubDate>Tue, 27 Jun 2017 17:04:43 GMT</pubDate>
            <atom:updated>2017-06-27T23:24:05.550Z</atom:updated>
            <content:encoded><![CDATA[<h3>What does local demand data bring to the table?</h3><figure><img alt="" src="https://cdn-images-1.medium.com/max/960/1*jO6-O4SRECfNIXcRvnlYVA.jpeg" /></figure><p>At IdealSpot we often speak about how important understanding local demand is for your business. But how does demand compare to the data you typically use in site selection models like income or spending?</p><p>Let’s approach this question via a comparison of spending and demand. Spending data estimates which consumers are spending money on the kinds of products you sell. Spending data is aggregated by the Bureau of Labor Statistics and distributed yearly via the BLS Consumer Expenditure Survey. Spending data is great for seeing the current state of the market but it does not quite demonstrate demand because consumers can only spend money where there are existing establishments (online sales excluded). It cannot tell you where the people who want to spend money on your product or service live, work, and play.</p><p>Demand, on the other hand, does capture this information. IdealSpot collects this information through social media platforms and search engines to get to the exact location where demand is being expressed even though there may be no establishments serving the desired product where demand is being expressed. If there are not enough providers to meet the demand, then the difference between spending and demand is profit for anyone able to capitalize on the open opportunity.</p><p>The below screenshots demonstrate this difference. Several zipcodes in Chicago have spending for glasses and optometrists but relatively low demand. However, some of the areas with low spending have high demand. What’s going on there?</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*3lUPwPqVY_j-GadJE8bXlw.png" /><figcaption>Spending for glasses and optometrists in Chicago by zipcode</figcaption></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*BJ7qLCPZBqvAusMZairDkw.png" /><figcaption>Demand for glasses and optometrists in Chicago by zipcode.</figcaption></figure><p>According to our above theory, the areas with low spending and high demand should have few glasses shops and optometrists. Is that true? Let’s check the business counts in two cases:</p><ol><li>Northwest of downtown there is both high spending and high demand.</li><li>Southwest of downtown there is low spending and high demand.</li></ol><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*Za93MYK7RSzIExosr0gjGw.png" /><figcaption>Compare the counts of glasses vendors and optometrists per zipcode to spending and demand above</figcaption></figure><p>In case 1, there are indeed many glasses shops and optometrists.</p><p>In case 2, there are not many glasses shops or optometrists despite the larger demand.</p><p>Why the difference? We’d have to dig in more to be certain, but it seems that while the region in case 2 is more populous, case 1 is more affluent. Glasses are often considered fashion accessories so providers may be choosing to target wealthier individuals in their location decisions. However, case 2 would certainly be a viable option for many lower-end eyewear brands or optometrists looking for a wider pool of patients.</p><p>Demand is an important piece of the site selection puzzle and it can only be found at <a href="https://www.idealspot.com/">IdealSpot</a>. Our suggestion: begin your search with demand and narrow in on your desired demographics until you find the locations that are right for you!</p><p><a href="https://www.idealspot.com/comparing-spending-and-demand/"><em>This article was originally posted on IdealSpot.com.</em></a></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=c91c147b12d0" width="1" height="1" alt=""><hr><p><a href="https://medium.com/idealspot/at-idealspot-we-speak-a-lot-about-how-important-understanding-local-demand-is-for-your-business-c91c147b12d0">Comparing Spending and Demand in Site Selection</a> was originally published in <a href="https://medium.com/idealspot">IdealSpot</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Hard-earned advice for AI products]]></title>
            <link>https://medium.com/data-science/hard-earned-advice-for-ai-products-c320d900862e?source=rss-2971933d869e------2</link>
            <guid isPermaLink="false">https://medium.com/p/c320d900862e</guid>
            <category><![CDATA[artificial-intelligence]]></category>
            <category><![CDATA[product-design]]></category>
            <category><![CDATA[complexity]]></category>
            <category><![CDATA[product-development]]></category>
            <category><![CDATA[machine-learning]]></category>
            <dc:creator><![CDATA[Josh Winters]]></dc:creator>
            <pubDate>Sat, 06 May 2017 04:13:31 GMT</pubDate>
            <atom:updated>2021-04-10T20:53:00.644Z</atom:updated>
            <content:encoded><![CDATA[<p>AI products are hard. They require solutions to problems that are impossible for the human mind to comprehend while simultaneously operating on a shoestring budget and finicky, sometimes rickety infrastructure. They require organization of innumerable contextless fragments of data from across the web and forced compression of distantly related insights into greater generalization when individually those insights are doing everything they can to stay right where they are.</p><p>If a product conquers those challenges, it must learn how to fend for itself in an open market filled with mature, non-AI competitors as well as several young intelligent startups flush with all the hot new tech coming out of Stanford and Google. Every AI product wants the market to think they have the greatest value — the best data, the most accurate models — but the market becomes confused when two solutions disagree or use methods that have not been validated in practice. (Indeed, AI products often fail in practice thanks to their fallible creators). Even when everything else is running smoothly, the industry may already have a working solution that is not significantly improved by machine learning or the market is otherwise not ready for an AI solution. The product then must shift and dig deeper into core industry issues until it uncovers new value currently hidden by complexity. Those shifts can require big, reaching changes in the product’s backing organization, and not all teams are capable of executing or surviving a pivot.</p><p>That said, it’s an exciting space with seemingly endless opportunity. I’ll never claim to have all the answers, but here are a few lessons I’ve learned as Director of Product at <a href="https://www.idealspot.com/">IdealSpot</a> over the last year and a half.</p><ol><li>AI has no inherent value. In fact, AI is more difficult to sell than human labor even accounting for efficiency improvements. Models, as with any other product, mean nothing until you put them into a context that works within your business domain. Most managers, executives, and clients you encounter won’t have a clue what you’re describing or why it’s important unless you can describe in plain English how your product helps their bottom line. You might never mention the underlying models or technology while selling the product, and that’s okay. Keep the whole context of your work in mind and explain your value in stupid simple terms.</li><li>Know your domain and know your limits. You won’t walk into a decades-old industry you know nothing about and revolutionize it overnight with some brilliant machine learning application you hacked together. You won’t. It doesn’t work that way. Especially in old industries, complexity has been rendered invisible even if it is ever-present. Humans are well-trained in glossing over complexity with ignorance when they have a process that accomplishes the task at hand. Once those processes are established they become distrustful of any new solution that might happen to come along. Instead of providing a new process to solve the old problem, provide a new tool that helps automate the old process. Make life easier but don’t assume agency or superiority. You must earn the opportunity to shape the world around your new way of thinking.</li><li>Be opportunistic. If the team backing your product doesn’t have experience or a solid network in your target domain you need to be willing to take small steps through intermediary roles until you can finally get your hands on the right work. You can start from anywhere if you keep your elbows sharp and keep moving toward your goal.</li><li>Encapsulate the whole stack — not just machine learning talent, but also modern databases, scraping, and enough about distributed systems to set up and scale your product from scratch. Data in particular is key. If you don’t have a unique set of diverse data at your core you’ll have trouble achieving either the quality or flexibility your product needs to survive. Data drives productization.</li><li>Magic doesn’t sell. In most cases, the models are easy but proving value is extremely difficult. Even if you build excellent models that optimally provide answers to the world’s hardest business problems, clients won’t trust you until you’ve been sufficiently tested and proven at scale. If that’s the route you choose to go you are missing out on all the low-hanging fruit that doesn’t require perfection or even significant sophistication to achieve adoption.</li><li>Despite all the hubbub, the AI industry hasn’t evolved enough yet to have earned an inherent trust in its functionality. There’s a whole class of AI apps in the wild right now that are learning the hard way that computers still aren’t as smart as humans, and a whole generation of consumers who are learning to distrust AI apps thanks to failed or difficult solutions to simple problems.</li><li>Use machine learning to more effectively draw forth product value. I’ve found that right now machine learning fits well as grease in a UX — do the guesswork and fuzzy reasoning that your user would otherwise do in your app whenever possible. Fill in appropriate default values and settings, interpolate missing information, predict actions and discover trends in interaction data to inform a better user experience. Model the user, not the domain.</li><li>Think of your product as an AI even if it has few intelligent components. Your goal is to have your users interact with your product as fluidly as they speak with a human. Remember that communication does not require speech or text, only a consistent pattern language</li></ol><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=c320d900862e" width="1" height="1" alt=""><hr><p><a href="https://medium.com/data-science/hard-earned-advice-for-ai-products-c320d900862e">Hard-earned advice for AI products</a> was originally published in <a href="https://medium.com/data-science">TDS Archive</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
    </channel>
</rss>