Marketing automation: workflows, data and measurement

---

---

Navigate this post about Marketing Automation

1. What is Marketing Automation?
2. Who is it for?
3. How does data play a role in it?
4. How does it work for businesses?
5. How will it be used in future?
6. What can you expect from it?
7. A final word

This is an Introduction to Marketing Automation

Marketing Automation tools do not replace your marketing team.

Have you been to any virtual conferences lately, or sat in virtual boardrooms with your marketing teams? Then chances are you have heard the phrase “Marketing Automation”. In case the term is new to you, here is the basic outline.

Marketing is likely to hand you an unhealthy amount of repeat work. There is often a lot of data to prepare for reports, or for splitting your audience into segments. There is the dull job of swapping out parts of an email template. Even building and launching campaigns can be dull and repetitive. Often you will feel that you cannot do anything meaningful, because you are stuck in a loop of the same tasks.

This is where Marketing Automation comes in. It helps you get your daily repeat tasks done by software instead of by hand. That frees up your time. At last, you can get to the more meaningful work you have never been able to start.

Now go one step beyond your own time. Think about how much more a large marketing department can get done with it. More of the team can then do more meaningful work. That could have more of an impact on the business as a whole.

Disclaimer: Marketing Automation tools do not replace your marketing team. They are tools that help your team. Smart marketers can add them to their existing arsenal of marketing weaponry, so they can get more done.

This is who Marketing Automation is for

Anyone who does marketing can use it, whether a person, a team or a business. It suits those who are short of time, or who want to get more out of their marketing work.

Marketers at almost any level can use it. That makes it a useful tool for everyone from the most junior to the most senior. This is not because it might do parts of their job better than they could. It is because it can do more parts of their job, with less effort and in less time.

It also isn’t only for large firms or wealthy corporates. It applies to anyone from the newest day one startup to some of the most established big firms. Many firms can gain from it if the way they set it up fits their specific needs. The key words in that last sentence are “their specific needs”, because one size does not fit all.

Firms need to think hard about the problems they want automation to solve. Each needs to think about its industry and the wider setting it does business in. It also needs to think about where in the firm automation would fit best to solve those problems. It is not something you buy off the shelf, flip the switch, et voila, you are making money.

Example: 

A large corporate firm with a longer sales cycle would not need the same setup as a fast-paced eCommerce website. The climate in each is completely different.

Note: If you have dipped your toes into the marketing automation pool before without success, ask whether the setup was right for your use case. Chances are the solution was not the best fit for what you needed.

This is how Data plays a role in Marketing Automation

Data is key if Marketing Automation is to work well. As the old saying goes, “good input means good output”. In other words, “better data gives us a sounder starting point”. Good data is needed, but it does not guarantee good results.

Think of data as the heart of the whole setup. The heart does not work well when it is weak.

To be useful, data needs to be clean. It also needs to be correct. The variables you use need to be precise and suited to the automation you are setting up.

Take automation driven by leads. Leads are people or businesses who might buy from you. You have gathered their contact details, for example through a form. The leads data can cause problems if contact names are wrong or arrive in the wrong format. That can cause trouble if you run complex automations that rely on it. You might use it to feed lookalike audiences. An ad platform builds these by finding people like those in a group you have, such as your leads. You might also use it to run automated media campaigns with machine learning. The same is true for any other data point. The example applies to any setup.

This applies to more than the data that comes in or goes out. It also applies at the back end, where you store and process the data. That covers the schema and structure of the databases, data warehouses and data pipelines you set up for this. A schema sets which fields a database has and what type of data each one will accept. A data warehouse brings data from many sources together for analysis. A data pipeline moves data from one system to another and may process it on the way.

Again, data is of no use if it is stored or processed wrongly. A Phone Number that is processed and stored as First Name data by mistake is useless. It should be obvious, but you would be surprised how often this error is left unfixed.

Data often acts as a variable in the rules that trigger an automation. That is why it needs to be precise. If the data is wrong, the rules you set up to trigger it may never be met. Then the trigger would not fire, and your automation would not work as you meant it to.

Example: 

1. Setup

You have set up your data pipeline to take in form data from a web page or an app. By mistake, you set up your automation so it sends form field data to the wrong columns in the database. You have also set a trigger based on data you would expect to see in one of the database columns (“1. DO THIS when Field Value is X” or “2. DO THIS when data is added to Column X”).

2. Errors

The form data is processed wrongly (meaning it may be in the wrong format). On top of that, it is stored wrongly. 

Sometimes the database will not store any data at all. This can happen when the schema rejects data of a type it has not been set up to accept.

3. Problem

With the example triggers above, the automation may do one of two things. Trigger 1 may not fire at all, because it never gets the data it expects. Trigger 2 may fire, but the processed data would be wrong.

That means the data cannot be used as planned. In some cases, you cannot reformat it or process it again to make it fit for use.

4. Conclusion

You can see how this could be a showstopper, both for the setup and for the business as a whole. It is a common error. You would be surprised how often it is left unfixed.

This is how Marketing Automation is working for businesses

Marketing teams keep trying new things, so automation now shows up in many ways. Some setups take over data processing or daily tasks for the team.

Some setups build segmented customer lists automatically. Some write the content of a marketing email by code. Others swap ad content on the fly to match how users behave, or even place bids in ad auctions for you. There are plenty of setups out there to explore and test.

Two uses worth exploring are automated audience segments and remarketing campaigns. Remarketing shows ads to people who have been to your site or app before.

One favoured use is Facebook Dynamic Ads for eCommerce. With it, marketing teams can target users at any stage of the purchase funnel. The funnel is the set of steps people may take on the way to buying. Content that suits each user is then shown to them automatically.

This Facebook Ads setup can support eCommerce sales. But its effect on revenue needs to be measured by each business that uses it.

How does this implementation work? 

Say we have three users visiting the eCommerce platform. USER A, USER B, and USER C.

USER A, USER B, and USER C visit the eCommerce platform. USER A shows a weak signal of interest in buying (Browse 4 Product pages). USER B shows a stronger signal of interest (Add 1 Product to Wishlist). USER C shows a stronger signal still (Add 5 Products to Cart and Initiate Checkout). But none of them buys.

We then build an algorithm that sets the rules for segments in our automation flow. The rules are based on what each user does on the eCommerce platform. These actions suggest interest; they do not prove what a person intends. Ads are then shown to eligible users based on their segment. The content is tailored to each segment, based on its relative interest and its place in the conversion funnel.

1. If a user visits a page and does nothing else, put them in the “Awareness” segment. Serve ads that teach them about the brand or product, to help build interest.
2. If a user visits a page and adds something to their wishlist or cart, put them in the “Interest” segment. Serve ads with images and details of items like those on their wishlist, or of the items in their cart.
3. If a user starts a checkout but does not finish it, put them in the “Intent” segment. Serve ads with coupons or discounts on their cart items, to test whether an offer helps them finish the purchase. Leave out people who have bought, and limit repeat messages. Before you target anyone, check consent and whether the platform allows it.

You can adapt and test the algorithm and theory above for other campaign types (App-Installs, Leads Generation, Newsletter Signups and so on) or ad platforms. It is one way you could split an audience into segments. A funnel is one way to picture the journey, though people do not always move through its stages in order. If you need to, you could layer on a bit more complexity to suit your specific needs.

NOTE: This is a basic example of the kind of algorithm online shops use to automate their marketing. It relies on the ad platform to do most of the work, without much input or upkeep from the marketer.

More experienced specialists will build bespoke solutions (using your platforms, databases and so on). Or they will build on what the ad platform can already automate. These setups often bring in SMS, Email or Online Messenger Services as well as the ad platform.

The firms I work with own sensitive intellectual property. I have used this basic, generic example to protect it. For more detailed examples or case studies, reach out to me. I will put you in touch with leading automation experts.

How does this help the business?

The example above helps cut the time spent on segmenting audiences, and on making ads and keeping them up to date. Marketers no longer need to spend as much time fine-tuning. They can turn to more valuable work.

It also means marketers no longer need to spend as much time and brainpower on small, detailed campaign tweaks. Tasks can run on their own, but someone still needs to set them up and watch them. Unsupervised learning finds patterns in data without labelled examples. Those are training examples that come with a known answer or category. It does not remove the need for human goals, limits and checks. Fewer limits do not by themselves make learning better.

All in all, the business gets more done for the same effort, and more of that work is meaningful. Whether that improves revenue depends on the costs, the setup and the results.

This is how Marketing Automation will be used in future

Disclaimer: These are new things I would personally like to see, or things I think we will see more often. Do not take them as predictions. I do not have a crystal ball to see into the future.

Note: Yes, I want machine learning and AI to do the bulk of the work. Machines can do some marketing tasks faster than I could. In my own view, it would be silly not to let them try (with a helmet).

Segmentation

1. Audience labels, set by machine to work out which bracket each user should fall under, without a person having to step in.
2. Content interest segments, built by machine to rank which type of content (and which content itself) a user is most likely to engage with.

Integrations

1. Integrations and uses for physical locations (automatically sorting people who visit “Brick & Mortar” stores into segments, by what they do there).
2. AI that works more naturally with Online Messenger services.
3. Easier access for firms with smaller budgets (tools that need only a few clicks).

Media Planning, Buying & Reporting

1. Tools that tag a website automatically (again, with only a few clicks).
2. Reports produced automatically (with in-depth insights written by AI).
3. Media plans drawn up by machine (letting AI build plans from report data and audience segments).
4. Ads made and placed by machine (this ties into the segments and media plans above).

Media

1. Smarter placements and ad space for DOOH (digital out-of-home), the ad screens people see outside their homes.
2. Smarter links with ride-hailing services, and smarter ad placements and ad space there.
3. Using online audiences for DOOH screens.

This is what to expect from Marketing Automation

As I said before, Marketing Automation does not replace a good marketing team. It is a tool that can help your marketing do more with less and earn a better return on what you invest.

Marketing Automation does not replace a good marketing team.

With the right setup, you could expect to save time on daily tasks and money on wasted ad spend. Better leads and higher revenue are possible outcomes to test, not promises.

Your marketing can gain from it when it fits the problem and its costs are worth it. Your team will be even more useful. They will be able to do a lot more of the fancy innovative stuff they could not get to before. You will move your business into the ’20s. You could even put yourself into a leading place in your industry, or make your current lead stronger.

You can target ads using clearer signals. Keep in mind, though, that those signals are incomplete. You will have a good idea of who you are targeting, what they are interested in, and what content is likely to tempt them to engage. It could help make your messages more relevant. Even so, relevance, privacy and how often people see them still need care.

Simply put, better results are the aim, and measuring tells you whether you reached them.

This is the final word on Data and Marketing Automation

It is not necessarily a new topic, but it is one that is coming back and growing fast. There are lots of chances for automation to make life simpler for marketers. It can free up their time for more of the fun stuff. 

Firms can also gain a lot as their marketing team gets more done in less time. Above all, that should add to their bottom line.

I look forward to seeing new players in the market and what they roll out. It is always great to learn from the competition. In fact, it would be hard to improve this kind of work without different points of view.

You’re welcome to get in touch with me on LinkedIn or by Email if you need help with your digital marketing.

Thank you for taking the time to read my thoughts on Marketing Automation and Data. I’d like to hear your thoughts too. Have you tried to set it up yourself? Which uses have worked best for you, and how did you set them up?

TL;DR (Much Time + Many Efficiency = Very ROI)

The heading is a playful shorthand. Financial ROI still has to be worked out: take the full costs away from the return, then divide by those costs. The costs of tools, setup, staff and upkeep can outweigh the time you save.

Enjoyed this post? Please share it if you did so more people can see this.

Have friends? Would they enjoy this post? Share it with them! They’ll appreciate it and so will I.

Here’s what Google says about Marketing Automation