Automation in Marketing

With all the hype with ChatGPT, AI and Machine Learning, what you can and what you should automate matters now more than ever. From writing press releases, to auto sending emails and reporting on insights, you can automate almost every task available in marketing including creating ads and videos. But what should be automated and at what level matters. As a marketing agency helping with agencies navigate with their data needs we have a few automation ideas that can help your business and a few that can really harm.

Reporting

When should you have a report or dashboard that’s automated, versus doing a one off ad-hoc pull. If you are running your marketing budgets out of excel, that is a sure way to make mistakes. If you are spending hours on your monthly reports that is another place where you can be more effective. However, is there a point in automating a reporting that you would only need once or maybe twice in the next year? Probably not! Therefore, as a rule if you are using a report frequently or it is consistent automate away. As marketing analytics firm, we identify a need for repetitive insightful reporting as compared to a deep-dive ad-hoc analysis.

For example, if you are running budgets on multiple campaigns, and channels, it is best to automate it with a pacing dashboard other then managing it in excel. On the other hand, if your client asked why do we see an increase in CPC, you can do a deep dive and get to the root cause which could be competitor pressure, inflation, wrong keyword targeting type, keywords are not converting, poor quality score on landing page and so on.

Presentations

Many tools such as Datorama, or Einstein AI, and many others have tools to present insights into your data. Typically, these models are trained to showcase anomalies for example, you were expecting X revenue but got Y on this date. They use simple trending analysis to showcase difference in your metrics. You can also prompt it to answer specific questions that you may have. While, these seem great, in reality without understanding your own data you may make the wrong assumption.

For example, let’s say you are running search ads, and one of your keywords saw an increase in CPC, the AI tool can identify the trend and let you know to pay attention to it. Before you pause the keyword, you may want to make sure it was not a one off. Typically, when you start running campaigns your CPC is higher than a few days or weeks later, as Google’s or Microsoft’s AI tools configure your campaigns. Therefore, making rash decisions without the full understanding of your business may be wrong. As a marketing analytics consultant we advise clients to configure their reports only after they understand the commands of the tool but also their own internal data.

Nurture Campaigns

All marketing automation tools these days allow for automated emails being sent as part of nurture campaign, or drip campaign. They are based on logic that you can set-up. The problem that most marketers do is automate it without testing. You should test when to deploy, how many emails, different subject lines and logic works prior to automation. To test, start with mapping out your logic and possible testing scenarios, running them manually prior to automating to a bigger group. It also only makes sense to automate if you are having a large number of leads coming in.

For example, say you have about 100 leads coming in daily, for three possible different target markets, you first need to establish what email to send out first, should be find more about our product, a promotion or let us know what we can help with as messaging. You may also want to test a few different subject lines. Maybe you are wondering if longer emails work best or short with one call to action. You may also test what day is best to deploy, this really differs by industry and type of product. You may also need to test if it is better to send email 2 and 3, 5 days apart, 7 or 14. Hint, for most B2B business the sweet spot is 7-10 days, but that can differ based on messaging, type of product/service and industry. With so many things to test, if you had less than 10 leads a day, it may not be feasible, so recommendation is to look at historical email performance and take the best educated guess and monitor performance and test over time.

Segmentation

Targeting is key in marketing, after all no matter what promotion or creative you are running if you are selling golf clubs to a person that does not golf its irrelevant. Therefore, should you automate your targeting. You can establish look-alike targeting in most if not all digital marketing platforms, there are even specialized segmentation tools out there that will help you identify and establish your segment all automatically. While segmentation works for B2C marketers in some cases, it maybe much more difficult to find the right target market for your B2B marketing. Segmentation models have been around since 1960s if not earlier, you should have at least one segmentation model done to identify your perfect target market. You can then use that target market for your marketing campaigns. Automated or not, establishing your target audience is a must.

For example, you may use re-marketing campaigns for eCommerce to target specific products your website visitors seen. You can also send similar products as an email campaign that usually works great. For B2B, you may decide to build a segmentation model based on historical sales or opportunities in your pipeline, utilizing third party data appending tools. This one off models can provide you with enough data to figure out what conferences to attend, where to run ads and so on. How you go about it depends on your company, some may benefit from third party tools, others may need to build a segmentation model or third may simple do a simple analysis and use targeting options in their digital marketing and media buys.

Digital Marketing

There are many opportunities in digital marketing for automation from keyword recommendations to testing different creatives to even having a tool write your campaign headlines, descriptions and even creating images and videos. But what parts of your digital marketing campaigns make sense to automate.  As digital marketing agency, we review the suggested keyword lists, we make our own creatives and write original copy, however, we automate reporting, pacing, and full loop automation. We find that the best results is when you combine human’s creativity and understanding of the industry/product with automation for logic and faster results.

For example, if we were to manage a digital ads campaign for a mobile app, we will automate dashboard which will showcase not only performance but also let us know if we are pacing towards our budget and our download goals. We would also let the system pick the best combination of creative design from video to display ads to headline/description. However, we would have designers and copywriters create creatives, we would identify our own segment to go after based on data available from third party tools. We would also test different creatives and messaging, different platforms from TikTok to Google Ads, using internal and externally created creatives and so much more.

Mixing and matching what to automate versus what to do manual depends largely on the time it takes to do a task, the effectiveness of automation and your skill level to use tools correctly.

About Marketing Automation Services

Hiring a marketing automation service provider goes beyond using a marketing automation tool such as Marketo, Eloqua or HubSpot, it can be applied to all your marketing efforts to building media buys, PR and running social media. The decision to automate and what part is an important question that each marketing CMO or agency should be asking. Here is a quick checklist to help make a decision to automate or not?

  • Frequency of task: If it’s a one off it is better to be done manually as you can get more insight that way and the cost of automation will be too great. However, if it is the same task that gets done repeatedly on the daily or monthly basis automation makes sense.
  • Level of Product or Industry required: Do you need have a highly technical understanding of the product, service or industry, if the answer is yes, then automating creative and messaging is not feasible.
  • Availability of Automation Technology: Is the cost of technology available affordable? Is it easy to use or can a mistake be made? And is the technology reliable? Some tools simply search the web for answers and will not provide any unique content for your website to use, in the case of using ChatGPT for example to write your blogs for you. It could also be super expensive as some analytical AI tools cost over $10,000 a year, to provide insight that you can easily create yourself with the write trending tools.

These are just guidelines but at the end of the day, using both manual processes and automation is the key to a success of any business.

About StrategicDB:

StrategicDB is a data marketing analytics agency, specializing in reporting, digital marketing and data cleaning/migration services. Anything that is data related, we utilize both manual and automation tools to get the best of both worlds. We have automated pacing dashboards, monthly presentations, helped build segmentation and lead scoring models, automated processes from simple naming convention in campaigns to running eCommerce ads for businesses. As data scientist first and marketers second, we bridge the two to come up with the best solution for your business, contact us today to get a quote on your data project.