《Adobe:如何利用机器学习提升营销数据分析效率和效果(英文版)(17页).pdf》由会员分享,可在线阅读,更多相关《Adobe:如何利用机器学习提升营销数据分析效率和效果(英文版)(17页).pdf(17页珍藏版)》请在三个皮匠报告上搜索。
1、1Data,Insight,Action:Machine Learning&AI for Marketing Analytics.2The path to data maturity.Wave 1-Aspire:demonstrate quick wins.Wave 2-Mature:build a single customer view(SCV).Wave 3-Mature:implement data governance.Wave 4-Mature:ensure data quality.Wave 5-Industrialise:automate and scale.Wave 6-Re
2、alise:maximising data science.Machine learning and AI in action.Use case:increasing share of wallet and lifetime value.Use case:reducing customer churn.Use case:improved individual customer messaging.Use case:re-engage lapsed customers.Use case:customer conversion optimization.Use case:contextual re
3、cognition.How Adobe powers the switch from data-driven to AI-driven.System of data.System of insights.System of engagement.567931414151515Table of contents.3Its easy to see the potential in artificial intelligence(AI)and machine learning(ML)for data analysis.Your team can use these tools
4、to surface deeper insights,process more data in less time,and automate the repetitive manual work of data cleansing and preparation.Despite the possibilities,organisations are still struggling to successfully adopt AI and ML.A 2018 Gartner report predicted that 85%of AI projects would eventually fai
5、l.Five years later,the prediction has proven accurate.1In order to succeed with a data project,its important to start with your business goals in mind,and use technology as a means to these ends.Its easy to get caught up in pure technology and pure databut the business value has to be the primary dr
6、iver.We worked with leading data and analytics experts to create this guide to implementing AI and ML in marketing analytics.With the right program in place,you can:Increase efficiency for analytics teams Successfully complete the data initiatives youre accountable for Automate lower-value tasks Rea
7、ch the right people,in the right channel at the right time Increase customer retention Drive more revenueRead on to begin(or continue)your journey.4Achieving data maturity.Before you can begin an AI and data initiative,its important to establish your organisations current level of data maturity.Your
8、 maturity level will determine which business goals should be your AI initiatives focus.Asplen-Taylor describes data maturity as developing in five consecutive waves,which may overlap slightly but are distinct from one another.2 They are:This section will explore how the goals in each wave contribut
9、e to data maturity and,ultimately,positive business outcomes.WaveThemeGoal1AspireDemonstrate quick wins2MatureData governance,data quality,single customer view3IndustrialiseAutomate,scale,optimise4RealiseClearer voice of the customer,maximising data science5DifferentiateFrom data-driven to AI-driven
10、4When data projects are treated as purely technology projects,rather than focusing on delivering value to the business,they fail.Simon Asplen-Taylor,Author,CEO DataTick5Aspire:demonstrate quick wins.Its important to show the value of data transformation early in your transformation process.Quick win
11、s with provable benefits can help employees and leaders alike get on board with your initiative.In a hybrid world where customers interact and transact in both physical and digital environments,web analytics alone are no longer enough for marketing teams to have a full picture of customer behaviours
12、.One quick win project to undertake is making the move from web analytics to omnichannel analytics.Omnichannel analytics include offline and online data sources,enabling analysis of the entire customer journey,not just the parts that happen on your website.In the Adobe suite,this distinction is the
13、difference between Adobe Analytics(web-based)and Customer Journey Analytics(omnichannel).Success in an omnichannel world requires a unified and customer-centric approach,balancing personalised experiences with privacy.As a result,organisations must adopt privacy-by-design principles,earn customer tr
14、ust and respect through transparent data use,and create real value for their customers and business.But to make this happen,leaders must replace silos with collaboration.5Omnichannel analytics in action.TSB Bank wanted to get to know its customers better,but a siloed approach to data made it challen
15、ging to get a complete view.They needed to consolidate online and offline data,as well as data across online channels,to truly understand the customer journey and provide consistent personalisation.With Adobe Real-Time CDP,Adobe Customer Journey Analytics,and Adobe Journey Optimizer,TSB Bank has the
16、 platform to provide real-time personalisation for their customers.THE RESULTS:200%sales increase in just nine weeks,and a 400%increase in loan applications in the first year of implementation.And thanks to increased efficiency and automation,the bank estimated a 1,000,000 savings in marketing costs
17、 in 2020 alone.Neil C.Hughes,Tech Columnist and host of Tech Talks Daily6Mature:build a single customer view(SCV).Data consolidation,governance and quality assurance are essential for creating a consistent single customer view.Even with the most sophisticated technology available,we simply cant make
18、 good use of customer data if its not all in one place.The single customer view is required to enable predictive machine learning and AI-assisted decision making.Its also crucial for permission management,legal compliance and other issues of data governance and quality.While there are other solution
19、s for storing and accessing data like data warehouses,a real-time CDP is the only purpose-built solution for creating individual SCV profiles that can be activated in real timewhich is where the value of the data is realised.Extending existing data lake systems to provide the same services as a CDP
20、requires substantial new development.Acquiring a CDP with these functions already available will usually be much easier,faster,and less expensive than developing custom versions or buying and integrating separate components for each.David Raab,Founder CDP InstituteWhat is a CDP?The CDP Institute onl
21、y offers its RealCDP real-time certification for platforms with three capabilities:1.Packaged software:A CDP is a single solution with all the required functionality2.Persistent,unified customer database:A CDP creates a single source of truth for customer data3.Accessible to other systems:A true CDP
22、 should have ample connectors and/or an open API to work with other solutionsWhat Makes a CDP Unique?Unlike a data lake or warehouse,a CDP is purpose-built to process and provide access to customer data.In general,CDPs have greater flexibility,more marketing capabilities,more privacy controls,and a
23、more user-friendly interface than other data consolidation platforms.Data governance can best be defined as“a system of decision rights and accountabilities for information-related processes,executed according to agreed-upon models which describe who can take what actions with what information,and w
24、hen,under what circumstances,using what methods(Data Governance Institute).”67Data Governance in Action Here is a hypothetical example of how a retailer can implement data governance across channels:Imagine a fitness apparel retailer,“Luma,”that operates both online and in-person.They need a way to
25、manage customer data,comply with regulations and restrictions,and respect customer consent.With Adobe Experience Platform,Luma can set boundaries for where consumer data will be stored.They can define roles and grant permissions to access,read,and edit the data.The data custodian can then create an
26、automatically-enforced data usage policy within Adobe Experience Platform.With these safeguards in place,the marketing team can make the best use of customer data without risking a breach of trust.Theres no need to choose between a lengthy data governance review or the risk of violating a policyauto
27、mated data governance increases efficiency while ensuring compliance.Mature:implement data governance.Data governance is crucial for compliance with regulations,of course.But more to the point,the marketing team relies on governance to build trust with customers.Marketing to a customer without their
28、 consent,or using data in a manner they havent explicitly agreed to,can erode trust and potentially end the relationship.78A solution like Adobe Experience Platform(AEP)helps automate and streamline data governance,making it easier to manage permissions and stay compliant.AEP brings governance issue
29、s together into a single framework that can help track data provenance,ensure accuracy,and keep it consistent across the organisation.Its important to note that data governance is a people and process problem as well as a technology problem.Part of achieving data governance maturity is setting expec
30、tations and developing a culture of smart data usage.With the ever expanding multitude of data sources,online and offline,and strengthened privacy regulations in a cookieless future,data analysts will be challenged to provide insights and a 360-degree customer view.Analysts can meet the challenge by
31、 keeping close track of the sources of their data,how the data flows,and who owns it.Carefully,transparently and easily making this data available to users can regain trust and bring real added value.8Yves Mulkers,Data and Analytics Strategist9Mature:ensure data quality.The most sophisticated analyt
32、ics algorithms cant make up for low quality data.If your marketing teams email list is populated with ,or donald.duckdisneyland.org,theyre not likely to get good results.Part of data quality is having the data available to the people who need it.If your marketing teams database has 100,000 profiles,
33、but sales can only see 25,000 of them,sales is likely missing out on opportunities.AI and machine learning can help ensure the accuracy,availability and timeliness of your data.Algorithms can reconcile data and eliminate false entries.A real-time CDP can make data available across the organisation w
34、ithout compromising security.Where AI really shines is in timeliness,however.AI-enabled data platforms can keep data current in real time,continuously updating customer profiles as new data is generated.10Industrialise:automate and scale.According to a recent survey,data scientists spend 45%of their
35、 time on data preparation.3 That leaves precious little time for the actual analysis that delivers the insights used to drive business results.Its easy to see why automation is an essential step in reaching data maturity.AI and machine learning can take on tasks like data cleansing,eliminating dupli
36、cates,and creating profiles.The other major opportunity for automation is in self-service reporting.Sharing insights with the stakeholders in your organisation doesnt have to require hours(or days)of building dashboards.Imagine if your organisations marketing team could log into their own dashboard,
37、see the insights youve gleaned,and even submit their own queries,all without sending your team a single email.You can make the data safely available to more stakeholders,with no coding required,and free up your team to take on higher-value challenges.Scaling Personalisation in Action Personalising e
38、xperiences for two billion customers might seem like an impossible challenge.But The Coca-Cola Company is taking it on with Adobe Experience Platform.“We needed to have a platform that would deliver personalization that is on a very small individual scale where weve got these mom-and-pop sort of sto
39、refronts in smaller markets,”says Keith Bartig,Director of Precision Marketing Technologies at Coca-Cola.“And we have the huge restaurant chains and grocery stores in the United States,and everything in between.”In order to support this scale,Coca-Colas technology team focused on standardisation,wit
40、h a central technology team managing the requests and requirements from markets around the world.“Everything we did,we did with global scale in mind,”Bartig says.“So,if one region had something they wanted us to build,we would add it to our repository of components and capabilities,so other markets
41、could use it in the future.”Whether you have 200 customers or 2,000,000,000,automation and standardisation are essential for scaling.11Realise:maximising data science.For data analysts,this is where the fun(and business value)truly takes flight.You have a single customer view,with data governance ma
42、naged,and low-value tasks automated.Time to experiment,optimise,implement and repeat.Adobe Experience Platform(AEP)and Adobe Real-time Customer Data Platform gives analysts the freedom to experiment without risking compliance or data integrity.Low/no-code solutions also make it faster and easier to
43、start experimentingno need to search code depositories for that perfect query.A platform like AEP is also more scalable than having individual analysts finding and implementing their own code.AEP provides a standardised environment thats easier to onboard into and easier to operate.If you choose to
44、use heavy code solutions,you will need developers to generate your data.However,developers can often be a limited resource,so you risk being too slow in your decision-making and in your organisation.To stay ahead in todays fast-paced business environment,its essential to eliminate as many dependenci
45、es as possible in order to make faster,data-driven decisions.11Lars Skjoldby,Founder of the Danish digital marketing agency Skjoldby&Co12Machine learning and AI in action.The previous sections went through the steps your team can take to create a trustworthy data repository,automate low-value tasks,
46、and implement self-service reporting.Now your team is free to explore some flagship use cases for machine learning and AI that can help lead you to better business outcomes.Use the full knowledge of your customers engagement history to cross-reference with similar customer profiles.This will create
47、cross-selling and upselling opportunities for complementary product suggestions,and will enrich the customer relationship from one-time,to lifetime.Those who go the extra mile can truly differentiate.If your analytics team can geolocate the customer and track weather conditions,they can promote loca
48、tion specific products.Tools like Adobe Realtime CDP and Adobe Journey Optimizer allow your marketing team to push the right offers to customers in the right channel.Most importantly,they can do all of the above automatically and at scale.Use case:increasing share of wallet and lifetime value.Market
49、ing teams can combine detailed audience segmentation,behavioral and demographic analytics,and large language model tools to personalize customer conversations and provide more intelligent,human-like customer support.For example,large language models can understand the context of customer inquiries a
50、nd conversation to provide relevant recommendations and answers and 24/7 support.This reduces customer wait times and enables human agents to address complex inquiries or issues,for an all-around better,more engaging customer experience.Ronald Van Loon,Principal Analyst and CEO,The Intelligent World
51、1213Use case:reducing customer churn.Use case:improved individual customer messaging.Use case:re-engaging lapsed customers.AI analysis can help spot customers who are at risk of churn earlier in the cycle and re-engage them with personalised offers.AI can also identify trends and anomalies over time
52、.This makes it easier to see where customers are encountering a problem that makes churn more likely.For example,AI analysis might show that customers who visit a specific page of your website are more likely to never buy again.This narrows down one problem to either the contents of that page,or the
53、 underlying issue that drives customers to visit it.While customers demand personalised content,they are increasingly opting out of sharing personal information.Fortunately,AI makes it possible to personalise even to anonymous users,by creating micro-segments based on behaviour rather than personal
54、information.For recognized users,AI can enable in-the-moment personalisation,delivered in real-time within a single micro-moment,from insight to action.When customers engagement level drops,its important for marketing to rekindle the relationship before the customer falls away completely.AI analysis
55、 can detect customers who are currently at risk,and over time will be able to identify them even earlier in the cycle.Your marketing team can use these insights to create personalised experiences for at-risk customers,sparking new engagement and higher lifetime value.14Use case:customer conversion o
56、ptimization.Use case:contextual recognition.Marketers currently make decisions about conversion by looking at an aggregated conversion rate.AI-assisted analytics can add intelligence and specificity to the process,enabling marketers to consider each customers individual needs.With your analysis on b
57、oard,marketers can create automated campaigns to reach customers with much more targeted and relevant content,thereby increasing the rate of conversion and keeping customers moving to a purchase decision.Timing and context are essential parts of successful marketing.Real-time personalisation is only
58、 possible through AI and machine learningit means operating in microseconds,far faster than a human could spot a trend and respond.With omnichannel data and SVC profiles,marketing teams can deliver real-time personalised experiences and content in any channel,regardless of the authentication state o
59、f the customer.Machine learning and AI can make humans more impactful to businesses by increasing efficiency and productivity,improving decision-making,enhancing customer experiences,and enabling better risk management.By automating routine tasks,analyzing large amounts of data,personalizing custome
60、r experiences,and identifying potential risks and opportunities,these technologies can help humans work more efficiently and effectively.However,its important to remember that these technologies are tools,and humans still need to be involved in using them effectively and making decisions based on th
61、eir insights.14Bernard Marr,Futurist and best-selling author15How Adobe powers the switch from data-driven to AI-driven.Data initiatives fail when they are focused solely on technology,rather than on achieving business goals.In other words,its essential to look at the bigger picture of how data beco
62、mes revenue.This process of extracting value from your data involves three interconnected systems for consolidation,analysis,insight and activation:System of Data:A platform to unite customer data across channels in real time to create actionable customer profiles.System of Insights:AI-powered data
63、unification and predictive analytics can interpret behaviour across channels.These insights feed back into the System of Customer Data to further enrich the customer profile and audience segmentation.System of Engagement:This system empowers marketing to act on your insights by orchestrating persona
64、lized journeys to deliver the right content to the right customer,on the right channel,at the right time.Marketing analytics is a very rich application area for business data analysts.From trend analysis,to customer behaviour analytics,to demand forecasting,to attribution analysis,to conversion and
65、retention modelling,there are many appealing opportunities for the analyst to gain insights from the flood of incoming business data and to make a significant contribution to the organisation.One sure-fire way to ignite these initiatives is to employ a data profiling tool that includes automatic dat
66、a quality assessments,outlier detection,cluster&segment analysis,trend&correlation detection,link&association discovery,and more.Such tools can uncover and guide the analyst to the most interesting,unexpected,novel,and engaging data insights.The added benefit of such tools is that they readily promo
67、te data democratisation across the business,even among those less data fluent,while fueling more efficient and effective data analytics tasks among the more data-fluent analysts.Kirk Borne,Founder and owner of Data Leadership Group LLC1516Adobe has developed a suite of solutions on the Adobe Experie
68、nce Platform(AEP)to help you achieve these goals.Heres how these solutions work together:System of Data:Adobe Real-Time Customer Data Platform(RTCDP)gives you privacy-ready Single Customer View(SCV)with consumer and account profiles that update automatically based on behavioural,transactional,and op
69、erational data.The data collected from across channels and systems is normalised into a standard taxonomy with person and account-level identity resolution,governance,segmentation,and activation in milliseconds.System of Insights:Adobe Customer Journey Analytics(CJA)analyses data across channels to
70、surface deeper customer behaviour insights both on and offline.It can collect and normalise omnichannel data including behavioural,transactional,and operational data.System of Engagement:Adobe Journey Optimizer(AJO)puts insights gathered from the CDP into action.Orchestrate and automate customer jou
71、rneys in response to real-time behaviour,contextual changes,or business signals.These intelligent solutions are trained on your organisations specific data before theyre put to work,and get smarter over time as new data comes in.Its often said that data is a business most valuable asset.But data wit
72、hout expert analysis is just taking up space.Data analysts can use AI and ML to uncover insights that drive higher retention rates,higher lifetime customer value,and much more.If youre ready to move from data-driven to AI-driven,request a personalised demo today.1617Adobe Experience PlatformAdobe Ex
73、perience Platform makes real-time customer experiences possible.As the foundation for Adobe Experience Cloud products and services,Experience Platform is an open system that stitches together customer data from every interaction through every channel in real time.The result is true,comprehensive cus
74、tomer profiles that drive relevant experiences for every customer.And it gives you the ability to analyze the data that really matters for customer experience,to train artificial intelligence and machine learning models that put your customers first,and to connect all your customer experience techno
75、logy to a single source of truth.Sources1.Gartner.(2018).Gartner Says Nearly Half of CIOs Are Planning to Deploy Artificial Intelligence.Retrieved from Gartner:https:/ and Analytics Strategy for Business:Unlock Data Assets and Increase Innovation with a Results-Driven Data Strategy.London:Kogan Page.https:/ 2023 Adobe.All rights reserved.Adobe,the Adobe logo,Adobe Real-time Customer Data Platform and Adobe Experience Platform are either registered trademarks or trademarks of Adobe in the United States and/or other countries.