Levels of Marketing Maturity: Examples of an AI-powered marketing strategy (advanced) and just setting up the basics (foundational)

Get ideas for implementing AI in your marketing organization (no matter how mature it is) when you join the AI Guild. Start your FREE trial at MECLABS.com/AI

Quick Case Study #1: How software company leveraged AI insight discovery and first-party data to decrease CPC from $44.15 to $13.19

dotData is a data science automation platform. The company wanted to determine the most effective and efficient segments to target for its ad campaigns.

BEFORE – Low-quality data and manual processes

The software company has a deep data set that includes numerous demographic, technographic, and behavioral characteristics for leads.

In other words, complexity – 100 data elements.

So the marketing team faced a significant hurdle – some of the most valuable insights and predictive signals would require a lot of complex data aggregations and transformations.

The team wanted to clean its data and find meaningful insights, especially from data patterns that are changing over time (weekly, monthly, quarterly, etc). They knew performing all these operations manually would not be practical, given the size of their marketing team and the deep statistical skill sets needed to analyze the data.

AFTER – AI-powered insights to discover audience segments

The marketing team decided to become a user of the company’s own technology, and leverage dotData’s AI-powered insights discovery technology (dotData Insight) to create segments for an advertising campaign on LinkedIn.

They built distinct audiences using their first-party data to determine which segments would have a high and low likelihood of converting and tested them against a baseline (‘neutral’) audience. “The two audiences (‘high’ and ‘low’) were both subsets of a retargeting list. So, this was a re-engagement campaign for leads who already existed. The ‘neutral’ audience was a cold audience using LinkedIn’s native targeting criteria,” said Walter Paliska, VP of Marketing, dotData. “The conversion event was downloading a case study from a lead gen form on LinkedIn.”

Of the 100 data elements, the artificial intelligence identified 25 signals as being important in predicting if a contact would convert in the campaign. For example, if the time between lead creation and the most recent website action was less than 12 days, there was a higher probability of conversion. Manually created leads had less chance of converting.

Here are some of the signals identified automatically by the AI technology that were most predictive of likeliness to convert on the ad:

• Time between lead creation and the most recent website action
• If prospect’s last engagement warranted updating the lead in the CRM
• The time needed to move a lead to the qualified stage
• If lead had been active on the website during the workweek
• Which campaign the prospect originated from
• Which product the prospect expressed interest in
• Job title

“The signals were specific to our data set and use case and should not be generalized as relevant signals for other use cases or companies,” Paliska cautioned.

Each prospect was given a score based on how well they fit each signal. The AI generated the weighting factors to determine the scores. The range of scores was from -325 to +276. Any prospect with a score above zero was put into the ‘high’ audience. Prospects with a negative score were put into the ‘low’ audience.

“By using our AI-powered technology, the work required to derive valuable data signals from raw data took only a few hours with minimal assistance from our data science team. We uploaded all three lists of in-house prospects to LinkedIn and placed them in identical content marketing campaigns for 90 days,” said Brandon Bednar, Digital Marketing Manager, dotData.

RESULTS – Lower CPM and CPC

Here’s how the how the ‘high’ and ‘low’ audiences performed:

• Clickthrough Rate – 2.19% for high, 0.73% for low
• Conversion Rate – 37% for high, 0% for low
• CPM – $288 for high, $509 for low
• CPC – $13.19 for high, $44.15 for low

The baseline ‘neutral’ audience performed worse than the ‘high’ audience and better than the ‘low’ audience with a 31.2% conversion rate and a 0.9% CTR. “This proved that narrowing our target audiences meant that we performed better than using LinkedIn’s approach,” Bednar said.

Quick Case Study #2: How new web design company got its first 63 visitors with SEO, social media marketing, and Udemy

The previous example’s team had plenty of first-party data to work with to optimize their advertising campaigns and segment their lists. In fact, many of the marketing examples we share on MarketingSherpa are from established companies further improving established marketing efforts.

But what if you’re just getting started? This next example is for a firm that had been around for only about a month and shows how it got its first few visitors in a very crowded and competitive space…along with some challenges it faced along the way.

SEO BEFORE – Lack of focus

Initially, Net Mafia’s on-page SEO lacked a specific focus keyword. The team relied on a basic page title, a simple meta description, and an easy-to-navigate URL slug. The page title and URL slug was just 'web design service page' and the meta description had only a simple CTA, 'Call us today to get your free web design service quotes.’

An initial SEO score evaluation using Yoast SEO plugin revealed areas for improvement, indicating the need for a more comprehensive keyword strategy.

Creative Sample #1: Search appearance evaluated in Yoast SEO

SEO AFTER – On-page SEO transformation

“Following comprehensive keyword research with tools like Ubersuggest, we revamped our on-page SEO strategy,” said Prerak Mehta, Founder, NetMafia.

The team listed approximately 100 keywords that they thought would be relevant to the brand and niche – for example, 'Web Design Company in Mumbai.’ Then, they exported the keywords list that immediate competitors use and reduced the initial list from 100 to 30 based on search difficulty (low to high) and search volume (high to low) to decide which keywords to focus on for each webpage.

“So I can take one keyword and put it in focus keyword in Yoast SEO, make sure that web page title starts with the keyword followed by additional text preferably describing the company, for example: 'Delhi Website Designing Company: NetMafia - #1 SEO Agency’,” Mehta said.

For the meta description, the team then expanded on the focus keyphrase ‘delhi website designing company’ – ‘Discover innovation with our Delhi website designing company. Seamlessly transition to an enhanced digital experience today!'

Creative Sample #2: Revamped search appearance evaluated in Yoast SEO with focus keyphrase ‘delhi website designing company’

SOCIAL MEDIA MARKETING BEFORE – Lack of focus

The team’s social media marketing efforts had a similar lack of focus. There was no consistent workflow, so they would have to create content every day.

The first 17 posts on their Instagram page were either carousel or single posts published on weekdays – five posts per working week, with one single or carousel post every weekday.

SOCIAL MEDIA MARKETING AFTER – Consistent workflow and more content, with an assist from AI

“After integrating Publer, we streamlined our social media workflow,” Mehta said. The team now allocates just one day to create and schedule content for the entire week.

Creative Sample #3: Content calendar screenshot

By streamlining the process, they were able to publish more content as well. “We increased our efforts and decided to post at least 10 reels, 3 single and 3 carousel posts on our Instagram page in a working week,” Mehta said.

In addition to the streamlined workflow, they used some other tools to help increase the amount of content. “Brandcrowd assisted us in taking readymade single and carousel post templates and making minor tweaks to match our brand style and content length,” Mehta said.

The team tried to use generative AI to help come up with new content ideas as well but were getting generic responses that wasted time until they got more specific. “To ensure that the content quality remained good, we gave ChatGPT extremely specific commands to help it grasp our business and marketing niche in order to target the proper audience and dream clients,” Mehta said.

They asked the AI chatbot to provide 100-200 topic ideas, and then inserted their own tips, tricks, and personally used tactics into an idea they liked, and instructed ChatGPT to use its copywriting abilities to combine the generic idea with the specific example.

AFTER – Trying to attract a previous audience

New startups often have to be scrappy and look for every opportunity they can to start building traffic and a following, even from their previous endeavors. In Mehta’s case, he saw an opportunity in the five courses he had created on Udemy since 2021 that had 35,910 total students. “During the COVID-19 pandemic, India where I live had a complete lockdown. Udemy courses were selling like crazy, and the algorithm was also pushing new instructor's courses more and more. Plus, my initial positive feedback and rating also helped in garnering such a large chunk of students,” Mehta said.

However, these students were not the team’s ideal clients – they were college students, not small business owners. And Mehta had been an engineering student himself, so some of the topics had nothing to do with his new venture. But others were somewhat related to – for example, building apps in C++, full stack web development, and a Figma beginner’s course.

“Redirecting them to my website may have increased the number of visitors, but if they jumped from my website to another or the average session duration was low, Google would eventually recognize my website content as lacking due to the high bounce rates and low engagement, resulting in a significant drop in my SEO score and Google page indexing,” Mehta said.

So he figured directing these students to the company’s Instagram page to get educational content would make more sense.

He sent an email with the subject line ‘Life Update: I have started my own company!’ that told students about his new endeavor and had his Instagram page as a CTA – ‘One simple way you can support me is by following my Instagram page, where I will be sharing insights, tips, and behind-the-scenes glimpses of my work.’

“To be honest, my email had no value proposition because I felt I was asking for a favor rather than selling anything,” he said.

Mehta couldn’t email them directly though because Udemy does not provide email addresses or allow the use of lead magnets to capture those addresses. So he had to send through the Udemy platform which doesn’t allow external links.

RESULTS – First few visitors, and a more streamlined approach

When it comes to social media, the team used to spend about one to two hours per day producing one post per day. But now they can complete the entire week's work in three to four hours (including three single, three carousel, and 10 reels posts with captions and hashtags). “Although it may appear to be a little complex, batch producing material is always easier and more efficient than daily. Not only my perspective, but if you check YouTube or contact other content providers, I am sure they will say the same thing,” Mehta advised.

In their first 28 days, they attracted 46 Instagram followers and 63 website visitors.

“We didn't use any PPC ads to gain visitors; everything was organic and social. Getting organic traffic takes time, preferably six to 12 months for freshly established websites, because even Google's web crawler requires a lot of data,” he said.

Related resources

Customer-First Marketing Strategy: The highest of the five levels of marketing maturity

Customer-First Marketing Guide: 4 steps (with case studies) to build a customer-first marketing strategy

Marketing Strategies: 3 proven customer engagement strategies to try (examples with results)