[{"data":1,"prerenderedAt":506},["ShallowReactive",2],{"blog-en-machine-learning-in-marketing-practice":3,"related-en-machine-learning-in-marketing-practice":274},{"id":4,"title":5,"author":6,"body":7,"category":258,"description":13,"extension":259,"image":260,"isToc":261,"langAlt":6,"meta":262,"metaDescription":6,"navigation":263,"path":264,"published":263,"publishedAt":265,"readingTimeMinutes":266,"readingTimeText":267,"relatedArticles":268,"seo":270,"stem":271,"teaser":272,"updatedAtCustom":6,"__hash__":273},"blog_en\u002Fen\u002Fblog\u002Fmachine-learning-in-marketing-practice.md","Machine Learning in Marketing Practice",null,{"type":8,"value":9,"toc":244},"minimark",[10,14,17,22,25,30,33,56,59,62,69,73,76,79,84,90,116,119,122,133,142,145,148,153,156,161,168,172,175,179,182,185,190,194,197,205,208,213,217,220,223,228,232,235,238],[11,12,13],"p",{},"While businesses are beginning to fully realize the potential of Machine Learning (ML), building ML models requires advanced data science skills and is a very tedious and time-consuming process. Automated ML focuses on improving the productivity of data scientists by recommending good ML models in a very short period of time and enabling data analysts, BI professionals, developers, and domain experts to build ML models without understanding the complexity of feature engineering, algorithm selection, and hyperparameter tuning. During a three-hour-long session, we highlighted several Machine Learning technologies and automated capabilities, to help understand the value it provides.",[11,15,16],{},"At the very beginning, we overviewed interesting and innovative technologies and applications of AI, mostly on the theoretical level, and gave a good introduction to the following practical presentations given by our Marketing Data Science experts.",[18,19,21],"h2",{"id":20},"part-1-azure-machine-learning","Part 1 | Azure Machine Learning",[11,23,24],{},"The first practical part of the workshop focused on Azure Machine Learning (AML). A short tool overview was followed by a business case of how to use ML to effectively motivate customers with discounts taking into consideration that some customers do not need to be given discounts, which in the end can bring significant savings.",[26,27,29],"h3",{"id":28},"introducing-azure-machine-learning","Introducing Azure Machine Learning",[11,31,32],{},"Azure Machine Learning is a tool developed by Microsoft providing a comprehensive solution for managing the whole process of creating machine learning models. This tool provides an option to create ML models for various professional levels:",[34,35,36,44,50],"ul",{},[37,38,39,43],"li",{},[40,41,42],"em",{},"Notebooks"," for data scientists who love coding in Python\u002FR",[37,45,46,49],{},[40,47,48],{},"Designer"," for analysts with knowledge of modeling process",[37,51,52,55],{},[40,53,54],{},"Automated ML"," option is useful especially for marketers and managers",[11,57,58],{},"It is possible to create models based on “standard” structured tabular data and also custom algorithms for text analytics and object recognition based on textual files or pictures. AML supports deployment of the models as a web service which is easily approachable via simple REST API or as a batch service for batch modeling within databases. Powerful computing ensures quick response, stability, and high accessibility.",[11,60,61],{},"AML covers the whole process of creating ML models:",[11,63,64],{},[65,66],"img",{"alt":67,"src":68},"","\u002Fupload\u002Fml-process.webp",[26,70,72],{"id":71},"business-case-optimizing-discount-popup","Business Case - Optimizing discount popup",[11,74,75],{},"One of our clients has an e-shop with apparel. They decided to run a campaign to motivate visitors to make a purchase on their website by offering them a 10% discount via a popup message when visiting the e-shop.",[11,77,78],{},"The current solution was not effective as it did not differentiate between various types of intents visitors could have and treated them equally. This led to a loss of interest resulting in loss of money!",[80,81,83],"h4",{"id":82},"more-efficient-solution-with-machine-learning","More efficient solution with Machine Learning",[11,85,86],{},[87,88,89],"strong",{},"Before:",[34,91,92,98,104],{},[37,93,94,97],{},[40,95,96],{},"Visitor A"," visited the website to see some inspiration and check offers. They were not interested in purchasing anything at that moment and they actually did not buy anything in the end (red cross). It would be more suitable to engage them via some inspirational content.",[37,99,100,103],{},[40,101,102],{},"Visitor B"," visited the web page after some investigation of its offer on product comparators, however, they were still hesitating. The discount stimulated them to purchase the shoes they always wanted.",[37,105,106,109,110,113],{},[40,107,108],{},"Visitor C"," visited the website with a clear intent to buy a specific product, they would buy regardless of the discount. In the end, they bought the product with a 10% discount. This is an additional cost for the e-shop.",[111,112],"br",{},[65,114],{"alt":67,"src":115},"\u002Fupload\u002Fbefore-ml-model.webp",[11,117,118],{},"The goal of the ML model is to make the popup content more efficient by predicting customers’ intent. To do so, visitors are divided into four groups based on the decision process funnel and the probability of their intent.",[11,120,121],{},"Each group receives personalized content in the popup. This solution can save money by not offering a discount to those visitors who would purchase the product anyway and moreover it moves down through the funnel all visitors that are in earlier stages of the decision process.",[11,123,124,125,128,129,132],{},"Visitors recognized as in the ",[40,126,127],{},"“Attention”"," stage with conversion probability below 25% are provided with inspirational content, increasing their engagement and moving them towards the ",[40,130,131],{},"“Interest”"," stage",[11,134,124,135,137,138,141],{},[40,136,131],{}," or “Desire” stage with conversion probability between 25% and 75% are offered a newsletter subscription which will tell them more about specific products and increase their motivation to move to the final ",[40,139,140],{},"“Action”"," stage.",[11,143,144],{},"Visitors recognized as in the “Action” stage with conversion probability between 75% and 90% are provided with a 10% discount voucher to motivate them to finish the conversion process.",[11,146,147],{},"Visitors recognized as in the “Action” stage with a conversion probability of over 90% are not provided with any offer, the model expects they would convert anyway, and no additional stimulation is needed.",[11,149,150],{},[65,151],{"alt":67,"src":152},"\u002Fupload\u002Ffunnel.webp",[11,154,155],{},"This particular model was trained on exemplary data of customers and their behavior on the web (from what channel they approach the website, what was the landing page, if they ever purchased something etc.).",[11,157,158],{},[65,159],{"alt":67,"src":160},"\u002Fupload\u002Fafter-ml-model.webp",[11,162,163,164,167],{},"After the deployment of the model, each user is provided with suited content leading to a ",[87,165,166],{},"30% profit increase"," and an escalated number of visitors moving through the decision process towards the action.",[18,169,171],{"id":170},"part-2-azure-databricks","Part 2 | Azure Databricks",[11,173,174],{},"The second practical part of the workshop concentrated on Azure Databricks. As an introduction, we went over the tool overview and continued with two business cases: effective product recommendation and optimizing smart bidding.",[26,176,178],{"id":177},"introducing-azure-databricks","Introducing Azure Databricks",[11,180,181],{},"Azure Databricks is an Azure service providing a very broad spectrum of possible uses. Serving people from many data-utilizing areas, it can process data from a wide range of sources and can output results into many other services. Incorporating the Apache Spark for data processing, Azure Databricks is constructed for simple and optimized use of big data.",[11,183,184],{},"Another main advantage of Databricks is allowing cooperation even at the same time and within the same notebook which is further enhanced by the support of multiple programming languages (Python, Scala, R, SQL, Java). The security is ensured by role-based access and integration of Azure Key Vault service for safe work with sensitive information. The usefulness of Azure Databricks is illustrated in two selected cases below.",[11,186,187],{},[65,188],{"alt":67,"src":189},"\u002Fupload\u002Fimage.webp",[26,191,193],{"id":192},"business-case-product-ranking","Business Case - Product ranking",[11,195,196],{},"Due to a generally high number of product pages, e-shops need to sort their products to ensure that customers do not need to search through many pages before finding the most favorite pieces. There are several ways to approach sorting which can be selected based on the company's needs and data availability.",[34,198,199,202],{},[37,200,201],{},"Sorting based on a single metric, like impressions and conversions would lead to fixing the same products on the top of the list and thus driving away bored customers. Therefore, when sorting based on product parameters, it is important to include other metrics, like past trends and profit margin to respond to trends and to optimize based on profit.",[37,203,204],{},"Alternatively, sorting can be personalized based on either related products to those a customer has already viewed or related customers by offering products they have shown interest in. Personalization may provide more tailored recommendations but requires significantly more data and more time and space for computation. By incorporating Azure Databricks into your solution you have the benefit of optimized big data processing along with the possibility to process data in real-time.",[11,206,207],{},"To evaluate the selected sorting of products, the A\u002FB test allows you to compare two groups of randomly divided customers which helps you to avoid interpreting the evaluation without mistaking the effect of improved sorting with unrelated external influence. If the first version of product ranking is successful (as is visible on the result from the A\u002FB test below), the continual improvement and subsequent testing can proceed.",[11,209,210],{},[65,211],{"alt":67,"src":212},"\u002Fupload\u002Fgraph.webp",[26,214,216],{"id":215},"business-case-improving-google-ads-smart-bidding-algorithm","Business Case - Improving Google Ads Smart Bidding Algorithm",[11,218,219],{},"The smart Bidding Algorithm for Google Ads is optimizing advertisement strategy by investing in ads that bring more profit. For this optimization, the algorithm needs to have correct and timely information about the profits from orders. However, this is often a problem e.g. due to returned orders or due to buying the products to store only after a customer has paid for them. For this reason, the Smart Bidding Algorithm calculates with revenues, assuming the simple constant margin and constant return rate for all products. This can be improved by predicting the profits - either using a complex model for the profit or separating profit into a multiplication of revenue, margin, and return rate where revenue is known, average margin can be taken from the database, and return rate can be predicted using a binary model.",[11,221,222],{},"Azure Databricks helps you combine the necessary data both from real-time streams and a regularly updated database. Predicted profits are loaded into Google Ads so that Smart Bidding Algorithm can improve the advertising in the browser and bring more customers and more future profit.",[11,224,225],{},[65,226],{"alt":67,"src":227},"\u002Fupload\u002Fgoogleads-smart-bidding.webp",[18,229,231],{"id":230},"summary","Summary",[11,233,234],{},"In summary, we showcased three practical uses of Machine Learning in e-Commerce, yet the number of ML applications is far bigger. With an increasing number of different tools providing automated algorithms it is always good to know what your data is flowing through before deciding on a specific option. With Azure, you can not only ensure security, but also it can be very easily connected to cloud storage and other tools that overall create one well-working and inter-connected environment, built for your convenience and, more importantly, increasing customer engagement and returns.",[11,236,237],{},"We have applied similar and many more solutions. Marketing combined with Data Science experience and technical expertise provides us with a competitive advantage to build custom solutions for each project.",[239,240,243],"action",{"link":241,"button":242},"\u002Fen\u002Fget-in-touch\u002F","Contact Us","\nLet us know, how we can help you grow.\n",{"title":67,"searchDepth":245,"depth":245,"links":246},2,[247,252,257],{"id":20,"depth":245,"text":21,"children":248},[249,251],{"id":28,"depth":250,"text":29},3,{"id":71,"depth":250,"text":72},{"id":170,"depth":245,"text":171,"children":253},[254,255,256],{"id":177,"depth":250,"text":178},{"id":192,"depth":250,"text":193},{"id":215,"depth":250,"text":216},{"id":230,"depth":245,"text":231},"Company","md","\u002Fupload\u002Fmachine-learning-marketing-banner.webp",false,{},true,"\u002Fen\u002Fblog\u002Fmachine-learning-in-marketing-practice","2020-10-07T01:49:55.000+00:00",8.2,"9 min read",[269],"content\u002Fen\u002Fblog\u002Fincrease-conversions-with-category-page-product-ranking.md",{"title":5,"description":13},"en\u002Fblog\u002Fmachine-learning-in-marketing-practice","A few months ago, we organized an online workshop on the practical use of Machine Learning. We talked about how to open new opportunities with Machine Learning in marketing as well as in other fields.","Sxt_6uoFUTDWOfF7FQ8d1O_IHFvxlepp4_lDBbuCfNU",[275],{"id":276,"title":277,"author":6,"body":278,"category":258,"description":282,"extension":259,"image":493,"isToc":261,"langAlt":6,"meta":494,"metaDescription":6,"navigation":263,"path":495,"published":263,"publishedAt":496,"readingTimeMinutes":497,"readingTimeText":498,"relatedArticles":499,"seo":502,"stem":503,"teaser":504,"updatedAtCustom":6,"__hash__":505},"blog_en\u002Fen\u002Fblog\u002Fincrease-conversions-with-category-page-product-ranking.md","Increase conversions with category page product ranking",{"type":8,"value":279,"toc":480},[280,283,290,294,297,301,304,308,311,315,318,322,325,330,334,337,341,344,348,351,355,358,362,365,369,372,376,380,383,387,390,394,397,445,447,450,467,470,477],[11,281,282],{},"Traditionally, e-commerce retailers and marketers pay most of their attention to developing product pages and checkout pages, because that is where the sales happen. However, we need to look at the previous steps, before the checkout. Customers use category pages to understand the offers of the e-shop, in other words, category pages become the attraction locations.",[11,284,285,286,289],{},"When a customer comes to an e-shop and looks for a product, category pages drive the majority of the site product discovery ranging between 50 - 70%, compared with other options, search results and recommendation sites bring around 10% each. This may seem like a lot and that there is not much to be improved. Nevertheless, t",[87,287,288],{},"he issue is in the progress from the category page."," When the category page is not compelling enough, customers are unlikely to reach the individual product page. Less than half of that traffic really proceeds to the product page. Customers probably did not find what they were looking for, the products were not relevant or outside their price range. By optimizing category pages, you can double the product discovery and increase the profit. Additionally, it builds a website structure, improves SEO, and subsequent remarketing advertising.",[18,291,293],{"id":292},"improving-category-pages","Improving category pages",[11,295,296],{},"Driving targeted traffic to category pages has been a topic of many marketers’ discussion. Commonly they have already adopted some improvements. The “science” behind the sales-driving category pages lies in displaying the optimal combination of selected products, counting on the limited number of the showed pieces. Here are just two examples, how usually e-shops try to tackle category pages:",[80,298,300],{"id":299},"manual-optimization","Manual optimization",[11,302,303],{},"The cooperation with many e-commerce businesses helped us understand that many e-shops are trying to optimize the category pages, however, they do it manually and rely on their own intuition rather than customer’s behavior. Manual arrangements can take days resulting in wasted resources; energy, time, and finances. The outcome of such activities costs more than they actually bring. Secondly, the changes cannot be applied fast enough to satisfy the customers’ needs.",[80,305,307],{"id":306},"category-page-ads","Category Page Ads",[11,309,310],{},"Targeting traffic via ads is another option of how to increase conversion rates. If done correctly it can bring a significant increase. On the other hand, the actual return on the investment is lower, taking the ad spent into consideration.",[18,312,314],{"id":313},"how-to-optimize-category-pages-deliver-better-results","How to optimize category pages & deliver better results",[11,316,317],{},"The optimization of category pages can be crucial in getting more website traffic, converting it to sales, and creating loyal customers from first-time shoppers. It is important to provide a valuable digital experience. When the category page doesn’t deliver what the customers expected, they leave without a purchase, not finding what they wanted. Relevancy is what matters. A crucial prerequisite to any calculation is historical data on products, sales, segments, etc. Without enough data, the results cannot be as satisfying.",[26,319,321],{"id":320},"personalization-and-product-recommendations","Personalization and product recommendations",[11,323,324],{},"The category pages need to be personalized to be able to achieve different goals for different segments. Adding a layer of personalization to different audiences’ levels up simple segmentation and yields higher returns. With first-time customers, you will probably focus on conversion rate while with loyal customers you can highlight a particular brand based on brand affinity, new products to complement already purchased ones, or something a little more diverse, depending on the customer profile. Assigning different products to each segment, based on the customer’s behavior on-site increases customer engagement. Tracking how the customer acts on the websites helps to understand their needs and display relevant items. If two people are looking for backpacks, they might be looking for a different kind. If one person is shopping for notepads, writing supplies, it is likely they will also need a school backpack. A different customer is looking at hiking boots and camping gear and might need a hiking backpack.",[11,326,327],{},[65,328],{"alt":67,"src":329},"\u002Fupload\u002Fproduct-recomendation-illustration.jpg",[80,331,333],{"id":332},"sorting","Sorting",[11,335,336],{},"Sorting products on the site in specific order or sequence based on their attributes, performance metrics, and their combination. The attributes can be price, size, brand, availability, etc. Metrics can be, for example, conversion rate, margin, revenue per impression, or inventory information. Attributes and metrics rely on the data about the products and the customers, collected from the website and internal databases. By adjusting the weights of the values, it is possible to promote and demote products in the sequence causing the relocation of the product on the page.",[80,338,340],{"id":339},"highlighting","Highlighting",[11,342,343],{},"Choosing to highlight specific products or groups of products, seasonal or campaign offers at the top of the category page supports marketing efforts. Placing some products on the most engaging and prominent spots on the sites creates a store-like experience. It is commonly used to promote new products, collaborations, and ranges. Highlighting can work for limited offers (discounts or weekend sales) and display products for a certain time period. Scheduling this should be aligned with marketing campaigns. Another option to adjust highlighting can be based on different business goals, chosen metrics, like profitability or liquidity.",[80,345,347],{"id":346},"segmentation","Segmentation",[11,349,350],{},"Segmenting your customers should happen on top of sorting and highlighting products. It allows creating category pages with specific product sequences that vary among different audiences. The marketing approach differs by different types of customers, their preferences, affinity, and different shopping stages, therefore category pages should be aligned with that as well.",[80,352,354],{"id":353},"personalization","Personalization",[11,356,357],{},"Showing the customers what they want to buy, the right time, the right product – like a personal shopper. It helps customers to discover items they really want. In e-commerce, personalization has a more significant impact and is an important part of modern shopping. It provides relevant recommendations for the particular customer segment (or with advanced algorithms, even down to each individual customer). Personalization creates an experience based on the customer’s behavior.",[18,359,361],{"id":360},"using-machine-learning-for-relevant-product-displaying","Using Machine Learning for relevant product displaying",[11,363,364],{},"Powering product recommendation with Machine learning allows you to be dynamic and automatically adapt to the changes. What customers want to see is what they really need, ideally on the very first page. Long searching is demotivating. Using historic purchases, similar customer behavior, or other factors can significantly change the way your customers interact with your e-shop.",[26,366,368],{"id":367},"product-ranking-based-on-customers-behavior-and-other-factors","Product ranking based on customers behavior and other factors",[11,370,371],{},"Product ranking in a sense of algorithms is a process of product scoring based on what the customers like. Additionally, the score can be evaluated based on the factors the e-shop defines, for example, storage availability (the more pieces of each product you have and need to sell, the higher score it gets). What we have found as an effective method is to look more deeply and find the optimal combination of the products display next to each other. Our procedure usually consists of data integration from various systems, followed by advanced analytics and machine learning algorithms for continual improvement.",[373,374,375],"note",{},"\nTo understand what product ranking is, think of it as product ranking from the system perspective, not as customer review. The ranking that you calculate helps the algorithm to show the right products on the category page.\n",[26,377,379],{"id":378},"testing-and-optimization","Testing and optimization",[11,381,382],{},"Finding an optimal solution isn’t an easy task, especially when personalization and merchandising are dynamic, evolving processes requiring repetition, calculation, and testing. AB Testing is a great approach for detecting the most profitable adjustments. It helps to discover the performance of different category pages or the effectiveness of product highlighting. It allows you to test competing strategies and make informed decisions for elevated results. It is possible to test a whole customer base or just a smaller part. With testing on different groups, you can experiment with bold ideas. You can see if sorting by high converting products is better than sorting by high-profit products. Or, you can assess if personalization delivers higher results than no personalization on category pages.",[26,384,386],{"id":385},"continuous-improvement","Continuous improvement",[11,388,389],{},"Once you set up the algorithms for relevant product displaying, you are not done yet. It is only a first step but not the last. Actually, the process is evolving, and you cannot stay static. Sometimes the results are oversimplified or too generalized, and you need to use more contextual data to improve the relevance. Thinking of it in the context of the customer journey helps to maintain the dynamics within the product associations.",[18,391,393],{"id":392},"what-to-think-about-before-diving-in","What to think about before diving in",[11,395,396],{},"Are you hooked yet? Ready to dive your e-commerce business into the ocean of personalization? Slow down a little bit. There are a few things to think about before you do any action.",[34,398,399,407,413,421,429,437],{},[37,400,401,404,406],{},[87,402,403],{},"Make sure your website is ready!",[111,405],{},"Some technology can decrease performance. Even if the personalization is great, it should not be implemented if the usability is diminished.",[37,408,409,410,412],{},"**Don’t change the entire website!",[111,411],{},"\n**Structural elements cannot be personalized, they should remain the same (cart, navigation panel, etc.)",[37,414,415,418,420],{},[87,416,417],{},"Less is more!",[111,419],{},"Too much of everything is confusing, too dynamic can be misleading. The key is not to look at personalization.",[37,422,423,426,428],{},[87,424,425],{},"Prepare your data!",[111,427],{},"Before you start the analyses and evaluations, gather all relevant data, the more historic data the better.",[37,430,431,434,436],{},[87,432,433],{},"Identify key factors!",[111,435],{},"Figure out, how you want to score\u002Frank\u002Frecommend the content. Do you need to clear your warehouse or promote more trendy items? What about seasonal stuff? Segmented or more individualized? More factors and segments, the more complicated and expensive it gets.",[37,438,439,442,444],{},[87,440,441],{},"Start small!",[111,443],{},"Try only a few changes first, test them, and then you can see if it is worth it to advance or not.",[18,446,231],{"id":230},[11,448,449],{},"Successful category pages drive performance and contribute to growing conversion and return ratios. Optimizing them generates a competitive advantage and brings multiple benefits to your e-commerce business and customers.",[34,451,452,455,458,461,464],{},[37,453,454],{},"Resource savings with automatization",[37,456,457],{},"Better performance with category pages optimization",[37,459,460],{},"Relevant experiences with personalization",[37,462,463],{},"Achieving more than one target with optimization and testing",[37,465,466],{},"Easy to scale with a growing product portfolio",[11,468,469],{},"It may seem super easy, just to set up a few rules and you are good to go. However, working with a large amount of data and continuous training of the algorithms is tricky. Rather than DIY everything, cooperating with more experienced professionals prevents the risk of “breaking it all” and losing your customers to it.",[11,471,472,473,476],{},"We have effectively set up the product ranking and relevant recommendations on category pages for many of our clients and helped them to ",[87,474,475],{},"achieve a 15-27% increase"," in the conversion only a few weeks after implementing the solution. We have been able to customize the solution based on different business needs and continuously improve the solution thanks to testing.",[239,478,479],{"link":241,"button":242},"\nCome to us today and help your customers to find the desired item tomorrow.\n",{"title":67,"searchDepth":245,"depth":245,"links":481},[482,483,486,491,492],{"id":292,"depth":245,"text":293},{"id":313,"depth":245,"text":314,"children":484},[485],{"id":320,"depth":250,"text":321},{"id":360,"depth":245,"text":361,"children":487},[488,489,490],{"id":367,"depth":250,"text":368},{"id":378,"depth":250,"text":379},{"id":385,"depth":250,"text":386},{"id":392,"depth":245,"text":393},{"id":230,"depth":245,"text":231},"\u002Fupload\u002Fcategoryranking-article-cover.webp",{},"\u002Fen\u002Fblog\u002Fincrease-conversions-with-category-page-product-ranking","2020-11-06T10:10:50.000+00:00",9.12,"10 min read",[500,501],"content\u002Fen\u002Fblog\u002Fgetting-the-most-out-of-permission-marketing.md","content\u002Fen\u002Fblog\u002Fmachine-learning-in-marketing-practice.md",{"title":277,"description":282},"en\u002Fblog\u002Fincrease-conversions-with-category-page-product-ranking","Online market is becoming very saturated and crowded with hundreds of e-shops, and the number is increasing. To keep up with the competition, every internet store must create a unique approach, provide enjoyable shopping and know its customers.","jo4l7QC10gq9zQLNdkKbEgxJz9GV44x0803c8Kv8BkM",1789131822919]