Leveraging Customer Engagement Data Insights for Powerful Retention and Engagement Journeys

Transcript

Hi, I’m Inga Romanov. Welcome to my session. I look forward to sharing some of the insights with you today. A little bit about myself. I definitely feel a part of the Marketo family at this point. I have founded a firm that specializes in Marketo consulting and a consultant myself. And I have also served as a New York Marketo user group leader since 2012. Today we’ll talk about customer data, finding it, looking at the different sources, synthesizing your insights, creating KPI that make a difference, mapping out customer segments and top priorities, and hyper-personalization at scale and all the tools available to you in Marketo. Then we’ll close off with takeaways. With engagement data, there’s really so much today. I know many of you are familiar with the funnel, but as we think about all the information available to us when we analyze engagement, there’s visitor data and understanding who comes to your website, who gets interested and engages with your content. Those who convert, MQLs, people who are ready to buy. Many of you are familiar with the traditional lead funnel, but in reality, there is a lot more data that’s available to us. Data from visitor data on the website, what do people respond to? What do they engage with? Lead data, those who convert it and were interested with some of the things you put out on the website as well as MQLs or even information from your sales team. Now the systems are much more complex and talk to each other. We have purchased data and conversion data from e-commerce or your fulfillment environment. So where do we start? You will also often in meetings will hear, but this drives traffic to the site. We can’t stop doing this. This is the most important thing. Maybe, maybe not. As an example, we had a client that changed leadership and for a few months their traffic dipped quite a bit. They hired a new person that did their SEO and SEM and you can clearly see the improvement in their traffic. It skyrocketed to be honest. We joined about three months later, but at the same time we started to hear a lot of complaints from sales about less qualified leads. They were drowning in students and as we analyzed the data, we found out that traffic was higher, but it was also less commercial, less informational. It was literally students finding the website based on that new keyword that the new person introduced and downloading the images. Traffic is great, but probably not the only thing that we will look at. It’s very important to listen to what is happening in the organization. Do you hear any feedback, complaints, issues or things that work really well? If we’re dealing with memberships, people are not renewing or maybe there’s lower conversion to memberships. Those are the types of things that you could attune to and include in your analysis. Understanding insights is very important. You’re probably going to work with so many metrics. I would recommend first starting looking at the available information and articulating what it is that you’re solving for. Shifting the paradigm from what I have to what I want. You will of course look across the channels, connect the dots, connect the insights, look at the type and fullness of data. Maybe you do have some information that’s very important, but the fullness might not be there. You might be only getting that information from one source. Consider who are your stakeholders in the company and how you will be communicating it to them. What is currently collected and already available to you, as well as data points that are critical but not available. Maybe this is something you can plan to collect over time. But only a few things that you might be looking at and this can get overwhelming very quickly. But besides our brand data and suggestions, recommendations of the different personas we’re targeting, there’s also customer engagement information starting from the traffic data in Google Analytics to marketer engagement data, email product specific, program specific, conversions or even other Salesforce data if you collect or if you enrich your data, as well as purchase information from e-commerce, from your fulfillment environment. With one company, for example, we had an engagement model that scored behind the log-in engagement with products that allowed for us to include that in our analysis as well. Membership, contract renewals, even exit surveys could become very useful for solving that one specific problem you’re looking to solve. You can also reach out to customer advisory panels or surveys with customer advisory, advisories to validate some of the theories you might be considering. And of course, speak to the sales teams. Although those are much more difficult to process their stories, they’re also useful to your overall analysis. There is a lot of data, but we also have better tools. We have AI that can really help you process some of the mass information as well as learn about new metrics if this is something you haven’t done before. I won’t have time to go over all of this, but I wanted to share a few pics for planning that you can use. In full disclosure, I use the paid account for all of this, but you could use ShareGPT for free. Now you set down, you analyze the data, you know what you’re solving for. And you’re ready to create actionable KPIs. You will be surprised analysis paralysis is much more common than you think. And it is our job to really focus on what we set out to do. As the next step, we’ll identify the most important audiences metrics, which will also track over time. KPIs is not one and done. Marketing programs will require ongoing optimizations. You will be able to set the baseline and check in regularly, depending on your company business lead lifecycle. It could be monthly, it could be quarterly, it could be on an annual basis, depending on the type of business. Of course, KPIs and this approach will not substitute your regular reporting, whether it’s you or your team members will continue to create detailed reports and continue to optimize within programs, within SEO, SEM. This is something that allows for you to solve for your bigger problems, more of a holistic approach, if you will. You can always adjust course if you are finding that engagement data as it becomes available over time is showing that we were not moving in the right direction, you can always adjust. And of course, a clear, concise snapshot of such performance is very key. Create transparency across the organization. Make sure you share it with stakeholders across, up, and to any people that may impact the outcomes. I wanted to provide a couple of examples. Here is a high level metrics dashboard for marketing. Context matters. Each of these, this is in Google Data Studio from different sources, each of these tabs has a specific context. Here from trends overall to for this particular company, engagement, retention, and conversions were very important. So we’re looking within those, within each of these tabs, we will look at specific data. And of course, for other more specialized teams, we’ll hand pick these KPIs and simplify the dashboard. So it’s more relevant to them. Here’s an example of a macro analysis of neutral engagement. With this company, we had many nurturers put together over years by different teams. And we wanted to understand from a high level, what works, what doesn’t, what do people engage with, are there overlaps? So we mapped all the nurturing programs and created these macro views of those to see. We filed some away, deprecated some of those. And in some cases, we had to refine the audiences because there were overlaps and people were getting very conflicting messages. Here’s another example of more specific KPIs and a baseline established. The goal was over one year period of time, improve conversions from lead to MQL as well as from lead to deals. Here’s how we mapped it out and recorded what we had at the very beginning so that we could look at it over time. I want to speak a little bit more about customer data and segments and how to approach that. Again, so much to choose from, depending on your type of business, B2B or B2C, you may have a lot of data available to you, including first party data provided by your prospects and customers. It’s very important to focus on parameters that matter to your company or relevant to products and services. Again, it might be something that you will simplify and go back to what you set out to do, what your specific goal is, what’s the problem you’re solving, is it membership renewals? So some of this in contextual reference, pick the parameters that are important to what you’re solving for. You might have a short term, so this is what’s available to me now, as well as long term in the next quarter or two, I would like to also add these data slices so I can look deeper in the performance. Here’s an example of a hard come butcher retailer. It’s a very simplistic prospect segmentation based on the lead lifecycle stage. In this case, content aligned to lead lifecycle stage, research, interest or decision, are they just heard about it or they’re ready to buy? But here’s an example of how taking the initial segments that you will focus on maps into Marketo. This is with an engagement stream in Marketo and people may move from stream to stream here to make this content more relevant to them. If you’re in B2B, it probably looks slightly different, but similarly your target segments will map directly to how it looks like in your chart. Focusing on key segments may seem in the beginning that we are ignoring everything else. What about the brand personas that we also want to target? However, in reality, this allows for you to reach these first priority segments better, have a better view of the buyer, be more relevant with your content, which typically yields more effective marketing strategy and much more meaningful experience for your customers. It drives retention, it drives conversions, and as you master each of these key segments, you continue to move on and expand your audience. In Marketo, we have a lot of tools available to us. You can create powerful personalization by using enhanced program types like engagement program versus default program, up your token game from person level tokens to program level tokens to velocity scripting or lead scoring. In Marketo, there’s no limit to lead scoring and you can also use it for product interest, implicit or explicit. You can use web personalization or Adobe Dynamic Chat conversations assisted with generative AI. Here, I wanted to go over a couple of examples. For instance, with engagement program, it’s highly suitable for multi-segment, multi-stream nurturing. You can take a very complex matrix and map it directly. And I often say it feels as if the engagement program was designed by an engineer who understands content very well. It’s easy to visualize your segments, map the content to your target audience and create transitions. The nurturing program has automated flows and transitions and scalable even for very complex buyer journeys. One of my favorite features is not only creating this program, but monitoring performance and improving content because it’s not always that the content is bad. It might not be right for this particular segment. It’s very easy to do in real time with nurturing program versus with default program while you have more flexibility and you can combine modalities. It’s definitely an excellent tool for complex logic and if this than that, but probably wouldn’t be as visual and easy to create multi-stream nurtures. Here’s another example of program tokens. I’m sure a lot of you are familiar with using them for variety of different programs. Remember, you can also use them for segment targeting that allows for you to eliminate extra email assets and target people with more relevant information. You can also set up UTM parameters at program level so you can track that performance later. I want to spend a little bit more time talking about velocity scripting. It is an incredible tool for hyper personalization anywhere from calculating a date to remind people about expiration of something to selecting values from objects. Let’s say if I have multiple subscriptions, it allows for me to send the email with a specific subscription name. You can print out one of several values in a table. For example, if it was an abandoned cart with several items I looked at. You can order content in the email based on specific criteria. I’ll give you an example in a second. Or show hide content based on user preferences. Another hyper personalization, footers in emails. It can also be done via velocity based on selection of the value in the language field. Here’s an example of such personalization with a recertification email. You can select a specific certification that needs to be renewed as well as format your date. Here’s an example with velocity where we have multiple products and we wanted to display them in this email in a very specific order. If they have one, two, three products or six products, note for each of the products there’s a module that includes image, article, article name, the teaser as well as the CTA link. We wanted to display them in a specific order which is strategically prioritized for us. Velocity scripting is excellent for that. Here’s another example. Your personalization data does not have to be from Marketo. It may be from another tool such as a recommendation engine. In this example, customer newsletter content, these blocks are personalized based on purchase products as well as web browsing behavior via another tool. As that data is updated in real time, the engine will send the recommendation to Marketo via custom object and you can use velocity to unfurl this in the email. I know not everybody is familiar with velocity but just remember there should be some velocity tokens for hyper personalization in your Marketo instance and you can just find a consultant who can help you with that. Don’t shy away. Another example, I mentioned lead scoring. In Marketo you can have multiple lead scoring models. They don’t all need to be lead lifecycle related. For example, in this particular setup, we had several credit card products and we wanted to send emails in the nurturing stream with priority that was both our internal strategic priorities as well as the interest that people showed implicit or explicit. Examples of implicit or explicit could be they visited the page, perhaps a pricing page with details about the credit card or they could have filled out a form with their interest requesting information. All of that can be placed in a score. In this case, I highly recommend keeping it as low as possible to make your workflows easier later. If you cap it at three, you can indicate low interest, medium interest or high interest and based on that rank order, nurture entry into one of these flows for these cards. I like using this graphic because it reminds us that if we’re working with Marketo, we’re likely in the center of all the data and tools that talk to each other. It’s easier for us to create a 360 degree approach because we’re not only specializing in one part versus another. To inspire you with takeaways, I’ll recap. Starting with a clear plan and mapping the specific goal you’re solving for. Looking at the data metrics and reports available to you to synthesize actionable insights. Don’t forget to try to use AI to organize thoughts or even learn about less familiar metrics. It will allow for you to create a more robust plan and move a lot faster. Increase visibility in the company with actionable KPIs. Make sure that you focus on specific top customer and prospect segments that you will target. You can always expand into other audiences later. Leverage the right Marketo tools to address your goals. There’s a lot of out of the box powerful hyper personalization tools available to you. And lastly, track performance over time, optimize continuously and change course if needed when you look at the engagement data. Thank you. This completes the formal part of the presentation. I look forward to your questions in the Q&A. And if you’d like to continue conversation later, join me on community and I’ll see you soon. Thank you, Inga. That was such a great session on translating the customer insights into the Marketo engaged programs. We have time for a 15 minute Q&A. So please drop your questions for Inga in the chat box below. Now let’s look at our first question. These reports are really helpful. I noticed that you don’t use Marketo engaged out of the box reporting. What tools or tech do you recommend small or medium businesses with limited budget to start with reporting? Thank you, Amy. What a great question. Can’t really say I’m not using Marketo out of the box reporting. Let’s just qualify these reports as an overall, if I’m trying to look at a problem I’m trying to solve, I need something else in addition to Marketo only. So if you’re a smaller budget, I think, well, not only smaller budget, but honestly, Google Data Studio is a great solution, which is often free unless you connect APIs. If you do want to connect different data sources, you’ll have to pay a fee, but it’s still probably cheaper than some of the more mature. Right now in the market, there’s a variety of great choices, but if I were to answer your question directly, go with Google Data Studio. Thanks for the recommendation there, Inga. The next question here is about velocity scripting. So this attendee wants to know, does Adobe offer any velocity scripting and training? Great. Well, velocity is, I mean, it’s a presentation and you really need to use it. It’s one of those where, forgive me for saying that, but don’t attempt this at home kind of a thing. Unless of course you have a long-term training plan for yourself and you would like to really learn it deeply, then you probably have to take a course like a Java so you understand the different objects. Generally speaking, look for solutions by Sandy Whiteman or myself, or just hire a consultant that you trust that can write a small snippet for you. Basically it’s not very expensive and it will get you right where you’re going without a huge investment of time on your end. But if you want to learn, there’s really no course for it on Adobe. There’s no course for velocity period. You really have to go and get down and dirty with programming, but just hire a consultant to write that code. And literally everybody on the session today, you should have sound velocity code in your instance. Great. Great. So the next question here is, how do you handle false email engagement in scoring models? For example, we have noticed when one person in an account takes an action in email, and all contest in that account are marked as taking the action. This is definitely an everyday kind of conversation. We have false positives for people showing opens when there aren’t any. We have false negatives. We have it both ways. We have, unfortunately, looking at email stats, not everything as it looks. We felt we could trust the clicks, but no longer with a very common use of email scanners. I don’t have a direct answer to this. Well, perhaps if they did, there would be Elon Musk. There would be somebody making a lot of money at this point. However, in a couple of instances, I’ve set up experimental smart listing triggers to mark leads that a person’s clearly clicking very many links in a short period of time. So this is something that you could try to do and track it for your instance specifically. Like I said, this is not something you can use. Here’s the solution. When we looked at the stats, we suppressed those experimental labeled as email scanners, and that really helps. However, when you look at email stats, mostly it makes sense over time and in comparison with others. Even if opens are not quite true, opens in comparison to others, opens over time, still makes sense. Thanks, Inga, for that answer. The next question is about activity logs. In Workout Engage, customers can use activity logs to understand all the records. How do you leverage that insights gathered through activity logs? Activity logs are fantastic. This is something that if you do have Salesforce and Salesforce admins on your team, teach them to go and look for that because they will just expand their knowledge and understand what’s really happening. It helps you troubleshoot and get down to complex problems quickly. However, activity logs typically apply on a person-by-person basis. If you have an issue with a thousand persons, it will be very difficult to go through a thousand different records looking, I mean, unless you’re AI, which I’m sure you’re not, if you’re on this session. However, you can use tools from consultants in our community that allow for you to download that data to analyze. However, if you’re in the instance, you’re human, you’re not downloading data, you’re not extracting, you’re not doing bulk extracts, it’s really only usable for a one-on-one basis. If you’re trying to get to the bottom of a problem, if you’re trying to get a trend and understand an issue, you have to make guesses. Somebody calls it, you have to make a lot of errors, mistakes to get to the right thing. I call it methodology. So you look for what could this be? Take one person, one record, look at their activity, find the things that you saw there might be a problem, and then you could create smart lists to find people who have that trend. And that should answer your question, unless you go to a solution that you have to pay for and then you can just download all the data and put it into queries either in Google Data Studio or any other data visualization you have. Thanks Inga for the suggestion. Here we have a comment from the attendee. I find sometimes personalization is interesting to do, but it really results in a statistically significant difference. So what would be your advice for this attendee to handle the strategy of personalization? This is such a great question. It’s a bit philosophical. Things changed a lot. Look at the last five, seven years, everybody consumer or B2B expect more personalization. They expect for you to know their preferences. They expect for you to know whether they want to be contacted or not. And while we’re still in some companies, here’s a request, an inquiry, and we call them. Not only it’s too expensive, it’s also not very effective. If someone is not ready, even if they have done some activity on your website, it may not result in a sale. They may be a detractor because of the fact that you reached out to them too early. So personalization and truly understanding of the different phases of the prospects, customers, if they’re buying multiple products is much more important than even two, three, five years ago. So personalization is the way to go, but how much you can do depends on how big your team is, how big your budget is, how much, I mean, honestly comes down to brain cycles. Do you have brain cycles to create that? Pick your battles, start with problems you can solve. You can crack on your own. If you’re a one-person team, find priorities, find things you can solve. Write down the long list of things you got to do and start with the ones you can solve on your own. Start with the ones you can solve in the short term and then plan for the long term once you solve those. Thank you for that recommendation, Inga. Now we are having a question back to velocity. This attendee said, I struggle with lead velocity. We allow people to MQL multiple times as well as other stages. This is because we have multiple products as well as a reroute. So it’s difficult to make velocity make sense. What are your recommendations for this attendee? Great. I love those types of questions and projects where it’s so complex that it hurts your head a little bit. Remember Marketo is a flat database. Velocity is only a tool to extract data from, let’s say custom objects or flat fields to display information. We do not have tools to use, well, it would be great, but we are where we are and we have tools that we’re working with. So if you’re looking to make sense and avoid errors, avoid issues with overlaps or conflicting nurtures, most of the times the data that is presented in relational customer database somewhere else, let’s say CRM, must be printed into the flat database. In other words, try to take the data from your custom fields, custom objects, and print them into the flat fields in Marketo. That will allow for you to control for entering the correct nurtures, making sure they’re not overlapped, removing people from nurtures when appropriate, routing persons from the correct lead partition to the correct partition. And that happens, it has to happen prior to entering Marketo. Marketo has particular strengths and you have to cater to that. And what happens to us marketers, MOPs, specialists is in companies we’re not always understood and we have to stand our ground in a way. We have to explain what Marketo is. It is not just an email tool, it’s a marketing automation platform, but it works in a specific way that we must support and feed correctly into that. Duplicating these controls, if you have custom objects with multiple purchases, if you have custom objects with different product information, brand information, make sure you’re also duplicating that data in flat fields, allowing for you to control for information and information flow and information outflow in Marketo. Thank you, Inga, for that suggestion. The next question is a bit more unique. So this customer is in the higher education industry. And the comment is, our engagement goals are different from B2C or B2B. We want to have high alumni engagement, even RSVPs, awareness and loyalty. What Marketo engaged tools would you prioritize given these tools? I love higher education. I know it’s complex. It’s a little bit matrix, perhaps, and this is where you might have difficulty creating segments, target audiences and sending the correct nurtures to people. It’s not only limited to your tools at hand. Unfortunately, when you have complex audience, and I’m literally speaking from my experience, just understanding this industry, just a bit of domain expertise here, you really need to sit down and it’s not the back of the napkin, but you get a large piece of paper. And I’m sorry for saying that, this is not a Marketo tool available to us. Get a large piece of paper or just draw it on the window, write out what do we do today? What are the audiences? What are the segments we’re targeting? What are the communications that are going out to them? Each one of them are working. What do we want to add? Put it in a different color. Once you see it in one place, it becomes very easy for you to look for specific metrics that will help you make decisions. You can drill down in Marketo, in Salesforce, in other tools. You might have some data visualization tool that’s available to you. Look for things that support you have specific questions. Do we have enough engagement? Are they converting? Are they coming back in person for in-person events? Look for all of those points of engagements, write it out, and that’s how you can create a model. If you had access to statisticians at your company, consider asking them, pull a file, consider asking them to write a regression model to understand those specific variables. Sorry, I’m getting a little mathematical. I’m going to stop the mathematical part. You don’t need that. You can, great. It does not matter. You can still map it out, look for specific parameters that made a difference, things that work versus things that didn’t work, and go from there. Thank you, Inga, for addressing that question. We now have a minute and a half, so I would love you to tackle this question as well, since it may resonate with a lot of audience. So how can we utilize MarketoEngaged standard reports with graphical representation for stakeholders? What are your thoughts on this? I love Marketo standard reports. I know there’s different opinions there, and you still have to run Marketo standard reports for most of your reporting. Visualization tools are great, but most of the questions you have, if you are in marketing operations, if you’re trying to solve a specific problem, you can solve with specific reports. One piece of advice, make sure you create contacts for what those reports are. Same report with people engaging with A versus B could look very differently if your definitions are not quite on point. Think through your audiences when you pull this. There’s different types of reports. There’s a person report, there’s email report, email click report. Each of those only makes sense within the context you’re pulling it. I know I had a whole slide on context, just like with AI context, and reporting your data out to others really matters. If you’re reporting on memberships, for example, this would be very different for audience only from those who are members as opposed to prospects or customers for other products. Look at the subsegments, make it granular. Think, they say measure seven times, cut once, do the same with your reports. Think through every audience and suppress things that don’t matter. For example, internal audiences, audiences that don’t matter for this segment. Look for ways to improve your audience before you report out. And yes, standard market reports are fantastic. They will answer a lot of questions you have. Thanks again to Inga for sharing all the great knowledge. I am now very motivated to try out your suggested applications to create hyper-personalized content.