Video Transcript

– [Ella] Thank you for coming to our session, and today we’re going to discuss the Campaign Mix. How to strike the perfect balance. I am Ella Aharon, a marketing data scientist in the professional services team.

– [Omer] And I’m Omer, research lab team leader, also from the professional services team.

– So, before we really dive in the Campaign Mix, I’m going to explain to you exactly what we mean by Campaign Mix. So, the Campaign Mix includes two types of campaigns. The first type of campaign is the cycle campaigns. So, the cycle campaigns are automated campaigns that you set up only one time and then they run on the recurrence that you choose, so either daily campaigns, weekly campaigns, bi-weekly, monthly, or even annual campaigns. Here are some examples of the cycle campaigns. So, you have reactivated customers on their eighth day that are in risk of churn, or you have new customers that are in their 11th day and have only one deposit or one purchase, and you have VIPs that are in risk of churn.

– The second type of campaign is the ad hoc campaign. Unlike the cycle campaigns, the ad hoc campaigns are executed manually in most of the cases related to real-life events. So, if you’re talking about the e-commerce industry, we can talk about Cyber Monday, Mother’s Day, or Black Friday. If we’re talking about the sports industry, it could be a Premier League match on Monday. It could be a Grand Slam final every year. If we’re talking about the casino industry, it could be new features, new games, new rooms, or even holidays.

Now, after understanding those two types of campaign and when trying to combine those two into one marketing plan, we might face two main challenges. The first challenge is the frequency. How many campaigns should we send to a specific target group in a specific timeframe making sure that we are not overexposing our customers? The second challenge is the prioritization. Which campaign would be better to send to each customer? If we need to send two campaigns at the same day, which campaign should we pick? Which campaign would be better? And I will start and talk about the frequency and in the next five minutes, I’m going to share with you one method to analyze your frequency. So, the first step is definition. Remember, our question is how many campaigns should we send to a specific target group in a specific timeframe? So, the first metric that we need to define is the target group. Which customers are we talking about? Are we talking about the active lifecycle stage? Are we talking about the new lifecycle stage? Those two target groups have different behavior, and this is why we need to be specific when choosing our target group.

The second metric that we need to define is the campaign channel. Are we talking about SMS? Are we talking about push, emails, or all of them together? It’s important to say that if we are executing 10,000 similar campaigns to 10,000 different customers, we’re basically talking about 10,000 different entities of campaigns. The third metric that we need to define is the timeframe. Again, our question is how many campaigns should we send to a specific target group in a specific timeframe? So, what is this timeframe? So, the timeframe is basically a sliding time interval which refer to the number of days before the campaign were sent, okay? So, it could be 3 days, 7 days, 10 days. Okay. It’s important to say that we are not talking about calendar timeframe. So, I won’t say Monday until Sunday, but I say seven days back since I sent the campaign. To give an example of all those definitions, our question could be, how many emails campaigns should we send to the active lifecycle stage in a timeframe of seven days?

The second step is mapping. And mapping is more of a technical step. We are measuring and mapping our database, our database of historical campaigns. The first thing that we need to define is the campaign type. In this example, you can see a single customer view, each line represents a day, and each cycle represent a campaign, in this case, an email campaign. You can see that the timeframe that we picked for this analysis is seven days’ duration, okay? The campaign type as you can see here in this example, this is a third type campaign. Why? Because it’s the third campaign that this customer got in the last seven days, okay? We’re taking the time back since we sent this campaign. So, it’s the third campaign. The type of this campaign is the third type campaign. This campaign, this is the timeframe, and this is the first campaign in the last seven days for this customer.

The second thing that we map is the response. The response could be measured by different KPIs. It could be activity days. It could be deposit days. It could be a payment day. It could even be an unsubscription, okay? In this example, the response rate is activity day, and you see that for each campaign, basically, we’re mapping whether the customer had a positive activity or a negative activity or no activity while he got this campaign. So, while we sent this campaign, this third campaign type, the customer had a positive response in this specific day. Then, we’re mapping all of our campaigns in our database.

The second thing on mapping and right after we are mapping our campaigns, it will be really easy for us to extract two main graphs. The first graph is the campaigns distribution. It’s pretty simple. The number of campaigns that we sent from each type. So, in this example, the active life cycle stage campaigns, 26% of the campaigns were third type campaigns. The third campaign in seven days. The second graph is the response rate, in this case, inactivity days for each type of the campaigns. And we can see a decrease for each type. We can see a sharp decrease after the fourth campaign which may give us a sign that maybe this should be the number that we need to define for this specific group, okay? Four campaign in seven days.

The third step is the analysis. In the analysis, basically, we’re running the same test over and over again, first, in different KPIs. So, in this example, payments’ day, inactivity days, or unsubscription, okay? Again, for each campaign type, we will get a different result. The second thing is to test it on a granular group. So if we talked about the active lifecycle stage, maybe we should divide it to active risk of churn and active no-risk of churn. Those two, again, different behavior for each group. If we’re talking about the new lifecycle stage, it could be the first week or second week. In this example, we tested on different KPIs. And as you can see in terms of number of payments, we can see a sharp decrease after the fourth campaign. In terms of activity days, again, a sharp decrease again, after the fourth campaign, and a sharp increase in terms of unsubscription, again, after the fourth campaign.

So now, we might think that this should be the limit for the active lifecycle stage, to send a maximum four campaigns in a week. And after defining the threshold and running more and more analysis, we’re defining the threshold for each group, segmented group, and placing the number. After we’re doing this, we need to prioritize between the campaigns. We know that the maximum of campaigns that we can send to each group in this example, four. Now, which campaign should I send? We need to prioritize between the campaigns, which lead us to the next challenge which is prioritization.

– Thank you. So, before I start talking about prioritization, I’m going to explain to you why exactly is it so important. So, prioritization is really important once you have one customer that is assigned to more than one target group. So, that’s when you really want to know if you only want to send only one campaign to one customer per day, then you really need to understand how to prioritize between the campaigns, so he will receive that campaign that you wanted him to receive and not a different campaign.

So, I think most of the CRM teams do prefer to prioritize their ad hoc campaigns in a higher priority. The question is, is it always the best solution? And I’m going to show you a few examples and practice with you, and just see a few situations where maybe it’s not the best solution to prioritize your ad hoc campaign in the top priority, and also, just show you a few dilemmas that might appear while doing so.

So, I’m going to start off with the e-commerce industry. So, you can see the blue campaigns are the cycle campaigns and they’re organized by the priority. So, in the top priority, we currently have the Churn cycle, right after that we have the Active cycle, after that we have the New cycle which is customers that have only one purchase, and maybe we want, you know, to make them have their second purchase. And in the top, sorry, in the bottom priority, we have our registered-only customers. They maybe only receive newsletter and we want to try and convert these customers.

Now, I’m going to add my ad hoc campaign. So, for the first example, I have my Black Friday promotion, and my Black Friday promotion I want to send it to all of my customers. Question is, where should I prioritize it? And I think, for this case, most of you will probably agree with me that we should prioritize it in the top priority. Why is that? Black Friday, huge deals, huge day. I mean, all of our customers know that it’s Black Friday today. They’ve got to get their promotions, you know. So, this is, I think, a fairly easy example of how to prioritize the ad hoc campaign.

But, what if, instead of Black Friday, we have a Mother’s Day? Not so important. Not as important as Black Friday. And let’s also add a birthday cycle. Birthday cycle, also an annual cycle. So, for example, let’s say I have a birthday today, and it’s also Mother’s Day today. What should I get? I mean, should I get my Mother’s Day promotion, or should I get my birthday promotion? And I can tell you that personally, I would prefer to get my birthday promotion. I would like to know that the company that I buy at know that today is my birthday. Maybe for me, Mother’s Day is not as important. So, this is, again, individual and this is only something that you might want to think about while prioritizing your cycle and ad hoc campaigns.

Now, let’s move on to the sport industry. So again, you can see the blue campaigns are the cycle campaigns. So, and of course, the prioritization is organized by the priority and these are only examples. So, we have our Churn cycle. These are churn customers that we want to reactivate them. We have our active customers that, yesterday, had a very big loss. And we have our Active cycle, all of the rest of the active customers. We have our New cycle and we have our Non-depositors that we’re trying to convert.

So now, I’m going to add my sports ad hoc campaign, and this will be the Spanish Clásico, all customers. So, Spanish Clásico, as you all know, is Barcelona and Madrid and Real, and it’s super-important game. You know, we want all of our customers to remember, watch this game, place a bet on this game. Super-important. So, I’m going to put it in the top priority, so all of our customers will remember to place a bet.

But now, is it the best solution? I mean, those active customers that yesterday had such a big loss, maybe it’s not the best way to send them a reminder to place another bet on a game. Maybe today, I want to compensate them on yesterday’s loss. You know, they want to know I know they lost yesterday and I want to give them something, so they won’t, you know, think they are always going to lose.

So, let’s just put the Spanish Clásico campaign right below the Active cycle, big loss yesterday campaign. So, this is great because our active customers will get the big loss yesterday if they had a big loss, and the rest of the active customers will get the Spanish Clásico reminder.

What is the problem now? Our churn customers. Our churn customers are in a higher priority now, so they will receive their usual cycle. Why is this a problem? Because we want them to receive the Spanish Clásico reminder. So, this brings me to my next point. My next point is something you can do with your ad hoc campaign is have it a bit more granular. And you can further segment your ad hoc campaigns into different segments. So, for this case, I’m going to divide my Spanish Clásico reminder into customers in the active lifecycle stage, and customers that are not in the active lifecycle stage.

So, here we have new churn and non-depositor’s lifecycle stages. So, let’s see. Churn customers, new customers, and non-depositor customers, they’re assigned to both of these campaigns and the Spanish Clásico campaigns. But since this isn’t the top priority, they will receive the reminder for the Clásico. Good. That’s exactly what we wanted for them to receive. And for the active customers, we have the active customers that yesterday had a big loss and they will receive exactly the big loss promotion. As far as the other active customers, they will receive their reminder for the Spanish Clásico, which exactly was our goal. So, by having your ad hoc campaigns a bit more granular, you can achieve two things. So, the first is you can send the campaigns only to the relevant customers. In this example, active cycle, big loss yesterday, we did not want them to receive the ad hoc campaigns, and they didn’t. So, this is good. The second thing we can achieve is to send different offers to different segments. And I’m going to show two examples of different offers to different segments.

So, the first one, I’m going to go back to my e-commerce and to my Black Friday promotion. Let’s say, we have our VIPs. Our VIPs, they’re good customers. We really want to, you know, give them a good offer. Let’s divide our Black Friday into VIPs and non-VIPs. And we do want to give them the same offer, but we want to give the VIPs three days’ duration, you know, they can use their promotion for three days, and other customers, non-VIPs, they’re going to have only one day. So, this is something that you can do if you further segment your ad hoc campaigns. If I move to the sport industry, for a different example, let’s say you want to have a reminder or a bonus reminder for a game for your new customers, your active customers, and your churn customers. Well, I guess churn customers behave differently than the new customers and the active customers because they’re not active, unlike the new and the active. You know, maybe they need a higher, you know, motivation to really reactivate. So, maybe I want to give them a different offer than the new and the active customers.

So, what I can do here is divide, again, the ad hoc campaigns into new and active, and churn. The churn will get a higher percentage of bonus or a higher, or a better promotion. And the new and the active customers will receive a less better promotion. And this way, we can achieve what we wanted to achieve. So, to wrap up, it is really important to understand your Campaign Mix, understand how to balance between your cycle campaigns and your ad hoc campaigns in order to deliver the most suitable campaigns to each one of your customers. And by doing that, you need to consider two things. The first is, you need to analyze what is your optimal frequency for the segmented group that you are targeting. And the second is, you need to prioritize your campaigns properly, so the most important campaign will be the one that your customer will receive and not a different campaign.