Speaker 0: Good evening, good morning, good afternoon, hello wherever you may be. Um, my name is Timo Kovala and, uh, I will be giving a session about how to optimize ad spend and ROAS with data cloud. Um, I am a lead architect, uh, at Capgemini. I work in The Nordics, meaning Finland, Sweden, Denmark, and Norway, and I’m responsible for, uh, Salesforce marketing and AI solutions. Um, first off, I want to say thank you to our organizers. Uh, it’s been a real pleasure preparing for this session and, uh, as this is my first time, I’m really excited and a bit nervous to start presenting to you. But hopefully, you will get some nuggets of insight from this session, and hopefully you will enjoy as much as I will. Without further ado, let’s dive into today’s topics.

So I had four topics in mind for today’s agenda. So first off, uh, we will be covering on a sort of general level what is ad spend, how you can take control of your ad spend. And then quite quickly, I’d like to move to, uh, data cloud because, as we all know, Salesforce is is the hot topic of today. Uh, so we want to be covering that quite quickly. Um, then, uh, I will be covering a topic that is a bit, uh, let’s say, um, off center with, uh, with data cloud, how you can get data out of data cloud and use that for ad segmenting. And then finally, we will be covering a few practical examples of, um, uh, data cloud driven ad segments that you can use, uh, in your work.

So I had four goals in mind after you have attended this session, what you would get in return. Um, first off, uh, I I think that, uh, after today’s session, you should be able to know, uh, what ad spend is, and then you should also know, uh, why why it’s important to care about ad spend. And then, of course, like, what is Data Cloud’s role in in ad spend, how it can help, uh, optimize ad spend, and finally, also how you can get started with using Data Cloud for ad spend optimization.

And, um, yeah. So today’s topic is about optimizing ad spend and ROAS, uh, in extension. So why are we so interested about ad spend? Ad spend is something that continues to rise in, uh, in companies globally. So, um, this year alone, uh, ad spend grew basically the fastest, uh, it has, uh, in in recent history. So 13% annual growth, um, on on ad spend, um, compared to previous year. And, uh, another interesting figure is that programmatic ad spend is growing three times as fast as non programmatic. Now programmatic is a technical term to basically describe ad spend that you or advertising that you buy with using technology like publishing platform or social media platform. So it’s advertising spend that is somehow technologically driven or bought. And non programmatic ad spend is pretty much like offline media or campaign media that, uh, you buy on on sort of ad hoc basis. And the third interesting figure is that, uh, on average, um, enterprises spend approximately 13% of their total revenue on on ad spend. So it’s a huge, uh, expense, uh, for enterprise customers. So this basically dwarfs anything that companies do, uh, on their, uh, sort of first party marketing. Uh, for example, email marketing, it doesn’t come anywhere near the, uh, numbers that companies use for, um, third party advertising. So it’s a huge, uh, expense. So, basically, a lot of marketing organizations are struggling how to get the most out of this, uh, expenditure.

Now what is actually ad spend, and what is return on ad spend? So ad spend is simply money that you use for uh, advertising like display banners, social media advertising, search engine advertising like Google Ads, uh, Bing Ads, and so on. And return on advertising spend measures, uh, basically this money that you spend on advertising times, uh, or or the revenue gained from advertising times the cost that you put for advertising. So very, very simple simple metric but difficult to optimize in practice.

Now on a general level, there’s roughly three ways to optimize advertising spend. So one is to optimize content. So that means basically when you’re doing, for example, social media posts or social media advertising, let’s say, um, you’re basically editing the taglines, uh, basically, for example, like, optimizing, um, the sort of caret that you’re presenting, like or or the call to action to get the most, uh, clicks to your website and so on or conversions. So editing the content, um, that you’re using on advertising. The second one is about cost optimization. So, basically, how to get the same revenue or same effect with, uh, lesser, um, amount of, uh, money paid per per clicks or cost per clicks or cost per impression, etcetera. And this is something that companies do quite a lot through agencies and also internally if they have, uh, the manpower to do so. But today, we are focusing on segment optimization. So that means basically, um, optimizing who you target. But as much as as you, uh, work on who you target, you should also optimize on who you don’t target with your, uh, advertising. So basically exclusions or, uh, disclusions of of of of of, um, marketing audiences.

So how where does StataCloud fit in, uh, in this picture? So how does Data Cloud optimize ROAS or help optimize it? Um, basically, there’s four things. So one, of course, is that, uh, because Data Cloud combines different data sources to build unified profiles of customers, It allows you to build much more targeted audiences that you would, for example, simply rely on CRM data or marketing automation data. Um, and this means that you get more refined lookalike audiences. So lookalikes, and we will be covering this shortly in more detail, uh, think of it like so that, uh, you have certain customers that you consider your ideal, uh, customers or ideal customer profiles that you basically want to duplicate that say that, hey. This is a good customer. I want more of these, please. So lookalike is basically, uh, identifying shared characteristics and building an audience based on on certain attribute or group of attributes that you find in existing data. And then third, and this is what I already, uh, mentioned a bit, is that you want to accurately exclude unwanted segments. So this may be, like, competitors, uh, customers who have recently purchased from you, customers who, uh, don’t work in the correct industry. For example, if you have a b to b context, uh, you might have different industries and and targeted advertising that you don’t want to target on the wrong industry, uh, companies. So, um, it’s it’s really great that you can actually, uh, pull in segments from basically the wrong customers and use those to, exclude social media or or search engine users based on, um, on your your data in data cloud. And, overall, uh, what is great about data cloud is that it it’s data that you own and and and it’s it’s first party data. So what it gives you is control. So you get control, uh, over over the audience, uh, audiences and and ad targeting as a result.

So, uh, looking at what capabilities Data Cloud has that are useful for advertising audiences, first off, we have unified profiles. So the ability to combine, for example, website data, um, CRM data, marketing automation data, uh, order management system, OMS data, uh, etcetera, to build these rich unified profiles that allow you to build much more targeted and holistic audiences. Sorry. Uh, then you have calculated insights. So let’s say that, uh, you have a custom grading system, uh, from a to f, uh, for for the most uh, promising to the least promising, uh, prospects that you can then use to build your segments in data cloud. Um, you could use calculated insights also for uh, classifying customers, say, to platinum, gold, and silver, and bronze customers, um, likelihood to convert, and why not also some sort of a churn risk evaluation? So, uh, calculated insights can be used, uh, to great extent, uh, in building these sort of very rich ad audiences. And then third, of course, like, it’s all nice that you can build these kind of profiles and and enrich them and build segments in data cloud, but you do need to get them out as well. So here’s where data activations come in. So uh, basically, they allow you to automatically push, um, segments with the most up to date data on a schedule, let’s say, once or twice a day to the desired ad platform. And this, uh, makes it pretty automatic to, uh, run this sort of ongoing ad campaigns.

And then a few ad audience best practices. So these are not, um, necessarily specific to data cloud, but, uh, advertising and building ad audiences in general. So what do you need, uh, when you’re building ad audiences? First off, and this is quite self evident, you need contact information. So if you want to match, let’s say, uh, CRM contact, uh, to a person who uses Facebook, you can’t use, uh, Salesforce ID for that matching. You need some sort of less unique parameters like pool name, email, mobile phone number, etcetera. And then Facebook, uh, can then try to match the Facebook users to those, uh, contacts that you provide via your ad audience. And this is why when you’re sending an ad audience from data cloud or some other system to, let’s say, Meta or Facebook, uh, you’re not going to get 100% coverage. Uh, it’s more like 50 to 70% coverage at most because people may have different email addresses they use for social media than than, uh, what you have in CRM system. Uh, then, uh, what is useful is, of course, because you can combine different datasets. You you have several user IDs, uh, stored in in, uh, data cloud, like mobile advertising ID or made. Uh, you may also include some sort of social platform IDs if you have those available, uh, and use those to identify segments, um, beforehand. So let’s say that if you’re pushing an audience to LinkedIn and you have gotten the LinkedIn, uh, user IDs for certain users, you might already sort of be able to, um, identify users who have that ID and then, uh, exclude the rest, uh, from the audience that you sent because those will likely not get matched, uh, when when you publish it to to link LinkedIn advertising. And then, uh, you can also include demographic data like age, birthday, uh, address, etcetera. On general, uh, you want to go for bigger audience sizes. So don’t, uh, do very niche, very small audiences, but aim for, uh, several thousand users per audience at minimum. And this, uh, is basically to ensure that, uh, both that you get, like, uh, the best uh, return on investment, uh, for your time building the audience, but also because a lot of these platforms have limitations that they don’t accept very small audiences. So that’s, uh, why you want to go for bigger ones. Uh, avoid gaps in the data. And here’s where you can leverage the unified profiles, um, feature. So, uh, if you don’t have certain, for example, contact point information in the CRM system, You can try to fill in the blanks using data that you get from marketing automation or some other system, uh, that is not necessarily integrated directly with Salesforce, uh, core platform. And finally, uh, when you’re building the segments, it’s fine to use very rich data, but don’t use that data, uh, in the payload that you are actually sending to external systems. Um, I don’t think that there’s necessarily, like, a a huge risk if you do accidentally send fields that are not needed. But, uh, on the safe side, it’s it’s good to only share data externally what you, uh, require specifically for the ad ad targeting.

Now for, uh, getting data out of data cloud is is a bit of a tricky thing compared to getting data in. So you know that data cloud has over 150 native connectors to get, uh, or ingest or access data. But, uh, getting data out of data cloud may require some additional work. So, basically, I’ve outlined six different ways that you can, uh, extract data from data cloud. One is data actions. So, uh, basically pushing, um, the data tables from data cloud directly to an external system or or even Salesforce platform. That’s one way. Uh, number two, and this is probably the preferred way if it’s, uh, if it’s possible for you to do. So you can use webhooks or, for example, marketing cloud, um, meaning marketing cloud engagement to actually activate the segments directly without any coding. And this allows you to use marketing cloud advertising, uh, when applicable for activating, um, ad segments if you have that, um, that, uh, platform in use. Then you can use data sharing. So let’s say that you have an external data lake like Snowflake or Azure or a Google Cloud Platform, um, you can actually, uh, share the or publish the ad segments from that, uh, repository, uh, to the advertising system. So this may be sometimes easier that way depending on, uh, your data pipeline, how it’s built up, and if there’s a ready made integration between the data lake and advertising platform. Then we have data graphs. So this is not exactly a method to push data out of data cloud, but it’s, uh, about, uh, combining different data tables into a single data table, uh, that you can easily share as a payload, uh, rather than trying to push customer data or audience data in several different batches to the target systems. So data graphs is like, uh, combining different data tables into, uh, one using a join, uh, inside data cloud. Then you have data cloud triggered flows. So, uh, this may not be such a great option, but, uh, it’s good to keep your options open. So in some cases, you might, uh, basically push the data actually via a, um, uh, outbound message from from a flow. But honestly, I I think that previous options are are the preferred ones. And and finally, you have MuleSoft. So if you’re using MuleSoft and have the APIs, uh, built already with that tool, um, then you can use that to export data as well.

Building an ad audience in data cloud pretty much follows the same structure as any segmentation. So you identify data sources, combine and unify data, build the segments, and then schedule an activation. So there’s nothing really strange about advertising in that sense. It follows the same same logical path as any segmentation in in data cloud.

And here are some practical examples of what kinds of audiences you can build with data cloud. So lookalikes, as said, it’s about copying your most interesting, your most, uh, active, your most profitable customers, let’s say. And you can build a lookalike, uh, already inside data cloud, uh, or you can push just the list of your most profitable customers or a subset of those, uh, to the external system, let’s say, Facebook Business Manager, and ask that system to create a lookalike audience for you. So, uh, depending on the situation and the size of the segment, you may choose, uh, either option. Then you can do an exclusion of, let’s say, recent buyers and and schedule that, uh, as a regular segment to the ad platform. You can identify cross sell or upsell opportunities uh, for current customers. And finally, you can use, uh, advertising as a way to win back, let’s say, unsubscribed or, uh, even churned or lost customers, uh, back to your customer base. So it’s basically one one option that you have left if you have, um, let’s say, if you’re limited in terms of regulation that you cannot, for example, due to GDPR, send emails or SMS messages to certain customers.

And finally, um, what are the three main benefits of using ad segmentation via Data Cloud? So number one is to be able to combine various data sources into a single umbrella and and use that to build rich audiences. Number two is uh, gaining less reliance on third party tracking like, uh, tracking pixels or third party cookies that are a bit of a problem these days. Uh, and uh, number three is that you get this kind of centralized command center of, uh, for for multichannel audiences. So let’s say that you want to target the same ad audience in several different unrelated platforms. What you get from Data Cloud is really the ability to orchestrate advertising across different platforms. So that’s a really powerful tool for any marketer.

And, uh, that’s it basically for today. So now I have time to cover a few questions. So, uh, let’s see what you have in in the chat. So there was a question at least from Bilal Bilal Ahmed. I hope I pronounced it rightly. Correct? Uh, so there’s, uh, uh, can advertising studio in Marketing Cloud use to optimize cost?

Speaker 0 (continuation): Good question. Uh, I don’t think that it’s it’s meant for cost optimization per se. So you can use it to optimize costs in, uh, via, um, exclusions, as I said. So don’t spend ads, uh, or don’t use ad spend on people that you shouldn’t be targeting. So in that sense, yes. But I don’t think that it’s a full fledged, like, uh, cost optimization platform. But really good good question. Any other questions? If not, then, uh, feel free to reach out anytime. Uh, connect via LinkedIn and I hope that you will have a really pleasant and interesting stay in Mar dreaming. It’s been a blast. Thanks for attending. Yes. Thank you for your great session. Thank you everybody for attending.