Marketing has long been thought of as the balance between Art and Science. But it’s only recently that we can maximize insights using advances in the sciences, such as mathematical modeling or machine learning, to truly respect the science side of marketing.
From Google Analytics to your marketing automation platform, there’s a goldmine of data just waiting for you to use. By taking big data and using mathematical approaches for analysis, businesses can maximize their efficiency and optimize performance while still focusing on inbound, outbound and account-based marketing approaches.
Here, let’s discuss how big data can complement your content, SEO, and website strategies to target the right people, with the right message, at the right time. With these opportunities, you can grow your audience, drive more meaningful interactions, and ultimately increase your ROI.
Turning Content Marketing from an Art into a Science
Content marketing is designed to be creative, to use words and visuals in new and compelling ways. What’s equally important is that your content marketing strategies reach your audience. This requires a well-thought-out strategy based on research and data.
In the past ten years, there has been a proliferation of tools available to create a data-driven strategy that take the guesswork out of the content process. Below are a few.
Google Trends
Using Google Trends, you can easily identify stories that are trending for your audience segment and investigate how to use that traffic to your advantage.
For example, you can use it to research cyclical keywords and plan your content strategy accordingly. It might not make sense to write a blog about the hottest vacation destinations in the middle of winter. If it’s the end of the year, you can imagine trending stories about predictions for the following year, such as this one on AI predictions for 2020. Right now, keyword research regarding COVID-19 is trending compared to other keywords, which can help us understand common patterns.
You can also use Google Trends to create topic clusters for your content. Topic clusters not only help shape your content strategy, but they will also help supercharge your SEO-focused content and make it easier to repurpose content on the same topic. Look for related queries in Google Trends to build out your topic clusters.
Social Media
Yes, social media serves a larger purpose than just being a place where you can post that perfect cappuccino art or video of your snoring dog. According to the social management platform, Hootsuite, nearly half the population has interacted with companies on at least one of their social platforms, and 41% say it’s important that companies they engage with have a strong presence. It’s not only another important channel for communication but also credibility.
Social media delivers insights that are useful for not only improving your brand’s social presence and performance but also understanding overall buyer sentiment.
Social Media Analytics
To improve social performance, spend some time looking at your social media analytics to gauge which content performs the best on each platform. Content that performs well on Instagram likely won’t perform the same on LinkedIn or Twitter. Cater your content to each platform based on how the audience likes to digest information. As WordStream explains:
“The one metric I really look at is comments per post. It tells me how engaged an audience is. No matter how much traffic you have, if you can’t cultivate an engaged audience you won’t be able to convert those visitors into customers.”
Not sure where to start with social media analytics tools? Check out this list from Buffer.
Social Listening
Beyond your brand’s owned social media pages, there are useful data points across the wider social web, specifically in tweets, posts, likes, hashtags, blogs, wikis, social opinion sites, and content sharing platforms like Slideshare and Youtube.
The importance of understanding the social pulse of your brand and topics that surround your brand have given rise to powerful social listening tools. Social listening tools allow you to mine and monitor the data on your prospect and customer’s thoughts to develop actionable insights. It’s like running a focus group that’s always on, gaining answers to difficult questions such as “what’s the overall brand sentiment of my competition?”, “what features do customers care the most about in our new product release?” and “who are our biggest brand advocates?”
With social listening tools, combining both qualitative and quantitative data points allows you to truly get inside your customer’s heads. Let’s say you have a brand podcast. Your analytics dashboard might provide the number of listeners, downloads, and playtime but social tools might provide the context, “what are people saying about it?”
The value of the insights from social listening analysis are far and wide across the marketing lifecycle, informing how well a new product release will go, giving a glimpse into competitor’s strengths and weaknesses, gauging customer satisfaction, providing sales leads, diminishing potential PR crisis matters, and providing real-time metrics on marketing campaign performance.
Buyer Personas
While not exactly a tool, buyer personas are the key to any data-driven content marketing strategy. For your content to perform well, it needs to be targeted to your ideal customers, or buyer personas. Using real-life data to craft your buyer personas will ensure that they are as accurate as possible. You can mine big data for your buyer personas from:
- Customer surveys
- In-person interviews
- Web/exit surveys
- Customer Relationship Management platforms
- Segments in Google Analytics
- Social media
- Email marketing statistics
- Past successes
Buyer personas include a combination of quantitative and qualitative attributes, such as age, occupation, hobbies, pain points/needs, and purchasing habits such as budget, timing, and where they go for their information. Buyer personas should be narrow, clear, and as identifiable as possible to make them actionable.
Once you have revenue data tied to your customers, go back and identify the content assets they engaged with throughout their journey and which keywords resonated with each persona. By knowing this type of information on past successes, you can further refine your brand’s buyer personas.
Building an accurate buyer persona will help you create engaging and useful content, target your ads more effectively, and nurture your ideal customers so they become advocates for your brand. The most effective digital strategies are the ones that consider multiple behavioral inputs, and then use those inputs to make predictions.
Optimizing the Relationship Between Big Data and Search
Google runs trillions of searches every year, and through their tools, there’s a wealth of data to understand user behavior.
By combining keyword data with insights from your brand’s digital assets, you can optimize both your web conversions and paid spend. Let’s look at a couple of approaches using common data analytics tools such as Google Analytics, Adobe Analytics Cloud, Tableau, SEMrush, and Webtrends.
Competitive Keyword Research
Keyword research plays an incredibly powerful role in informing your overall digital strategy. Google search queries use keywords to read and categorize your web presence accurately. Since search platform algorithms were built by parsing content into readable, quantifiable data, with the availability of new tools, it’s easier than ever for marketers to find insights on the search terms people use.
To optimize your ad spend for PPC campaigns, it’s critical to find the overlap of high traffic keywords, low paid competition, and high intent language. Intent language includes phrases like “where to buy,” “comparison,” “best,” and “top.” Intent keywords help you target more qualified visitors. Statistical methods, such as Logistic Regression, help marketers determine the keywords with the highest probability for engagement based on these factors.
Another approach is to run PPC campaigns on organic keywords that your competitors are ranking for and see if they convert for your brand. By doing so, not only will you potentially discover overlooked keywords, but you’ll also find ones that produce a higher ROI because there is already buyer intent.
This innocuous spying can be done using Google Trends, SEMRush, or a more advanced analytics tool like Webtrends. SEMRush allows you to see your competitor’s top-ranked SEO pages and conduct a competitor gap analysis, comparing each of your webpage’s performance side by side to identify areas for improvement. It also lets you not only see competitor’s SEM and display ad budget and key targeting attributes like audience and interests but also shows what their ads look like and what they say.
Similarly, Webtrends, an enterprise application, lets you take large volumes of data on user-based search properties for your brand and competitor organic keyword rankings, taking your analysis one step further. If you can’t beat them, join them.
Once you’ve identified a list of keywords to target for PPC, A/B split testing will help you quickly optimize ROI. The best way to A/B split test is to run two identical ad campaigns, except for a change to a single variable. Common variables to test include titles, meta descriptions, and calls-to-action.
By running these tests, you can quickly identify the campaign that resonates with your audience and allocate spend accordingly. The data will also help inform your landing page copy because you can determine which ads perform the best.
Using Existing Big Data To Optimize PPC
By analyzing your data, you can optimize your paid spend so no dollar is wasted on “blind testing.”
Adobe Marketing Cloud allows you to take your analytics practice one step further by distilling complex information into actionable visualizations of comparisons, trends, relationships, and more. Using their visualization interface, you can build dashboards to understand the relationship between multiple variables, such as which keywords converted and which brand touch was the most effective. You can also analyze your data to view advanced ROI metrics such as revenue per visit and revenue per conversion.
This level of context becomes increasingly important when deciding which keywords to invest more in and which ones to abandon.
Data analysis should not only be used to tell you what is working, but also what isn’t. Abnormalities in traffic patterns, behavior, and conversion patterns should be your first sign to drill deeper.
If there’s a drop in traffic, understanding why will help put you back on track. Did your rankings drop? Are fewer people searching for that particular keyword? Did a competitor decide to run a large PPC campaign to compete against you?
What if traffic remains high, but you see an uptick in conversions that aren’t within your target audience? Wouldn’t you want to learn the context of how people use a certain keyword? For example, if you sell to enterprise buyers but all of a sudden see a lot of students landing on your website and requesting demos, you might want to tweak your campaign on “free tools” to include the word “enterprise”.
With an advanced data analytics process called anomaly detection, marketers can get alerts on outliers that exist in the data, without going in and seeking them out one-by-one.
In addition to finding anomaly patterns, you can compare segments such as buyer personas and keyword/conversion data to discover non-obvious relationships. You may discover a certain persona, such as a CFO of a medium-sized company, converts when high-level quantitative language is present, while the CMO responds more to qualitative language based on real-world examples.
We recommend setting up segmentations that make sense for your business by finding and identifying the keywords that have generated the most engagement and conversions for your business. Creating segments can be done in both Google and Adobe.
With Content Grouping in Google Analytics, you can segment by format, content, or campaign to get an aggregate view of what is performing well and what isn’t. This will help you find abnormalities in traffic, behavior, or conversion patterns and make them easier to understand.
Personalize Your Website’s User Experience
In addition to optimizing paid spend, there are several opportunities to use big data to engage your audience in more meaningful ways on your website. With every visit, prospective customers provide you with copious amounts of information that can be used to understand their feelings. It’s then up to you to distill the data down to something actionable. Let’s take a look at a couple of examples.
Heat Mapping
Heat mapping software such as Hotjar and Lucky Orange are powerful tools that help you understand what your users do on your website, from clicking to scrolling to pausing and pondering. This information can help you easily improve your site’s customer experience.
Lucky Orange also allows you to use those insights to trigger 1:1 real-time alerts. Say a customer lingers on the pricing page, going back and forth between pricing plans. The platform allows you to automatically trigger a request to chat, such as “Questions on our plans? Let us know how we can help”.
Let’s take a look at an example of how this data can be used. Boston Consulting Group found that mobile usage influences more than 40% of revenue in leading B2B companies and 60% of B2B buyers said mobile played a significant role in a recent purchase. Heat mapping tools can help identify how users behave differently between desktop and mobile.
A/B Split Testing
Another opportunity to optimize your user’s web experience is A/B split testing. According to Forrester, better user experience (UX) design driven by A/B testing can increase a company’s conversion rate by 400%.
Four proven opportunities to improve conversions for A/B split testing include CTAs, headline text, headline images, and product features. By diving into data beyond click-through rates and conversions, you truly start to put yourself in the shoes of your customers. Doing so will allow you to drive more meaningful conversations by personalizing pop-ups, email workflows, and ad campaigns.
While traditional A/B web testing might tell you which has a higher rate of conversion, the combination of heat mapping and testing help you understand why.
Another indirect benefit of improving the overall user experience on your website is better keyword performance. Metrics such as click-through rate, bounce rate, time on site, pages per visit, and repeat visits all impact SEO performance and your PPC ad frequency.
Uncovering New Opportunities Using Machine Learning
Machine learning will take your ability to predict outcomes up a notch. While the large players, like Facebook, Netflix, and Amazon get the most headlines for it, smaller companies you probably have never heard of are also growing revenue by significant margins by leveraging the latest in AI for content personalization, customer insights, and beyond.
What is Machine Learning?
Machine learning is simply a subset of artificial intelligence. MIT Technology Review provides a simple and easy to understand definition without all the buzzwords:
“Machine-learning algorithms use statistics to find patterns in massive amounts of data. And data, here, encompasses a lot of things—numbers, words, images, clicks, what have you.”
One of the advantages of machine learning is its dynamic nature. As more data is fed into the system, the better the algorithm performs.
How Machine Learning Works
The fundamentals of machine learning can be boiled down to a two-step process. First, you take historical data and run mathematical models against it to identify patterns and predictions. Then, as new information comes in, you can feed that data back into the model to make it smarter and fuel new predictions based on the inputs.
What does this look like in practice? Let’s take a look at Facebook’s machine learning engine called FBLearner Flow. Facebook has built one of the most advanced predictive engines to figure out your preferences, including what you will buy, what you will read, what you will engage with, and more.
Here’s how it works:
- As a user’s information comes in, it’s fed into their machine learning tool.
- The tool runs a model against that data that has been trained on millions and millions of data points across users.
- Based on patterns, the model identifies a group of people that behave similarly to that user and puts them in a group.
- Facebook then allows brands to target groups based on this information. These groups identify individuals that are most likely going to engage with a company’s messaging.
Similar to Facebook, Netflix leverages machine learning to recommend which television or movies to watch. 80% of the content people watch on their platform is through their recommendation engine.
This level of personalization can also be applied to make clothing recommendations on an e-commerce site or content recommendations for a software company. According to the 2019 CMO Survey, the #1 marketing application of AI is content personalization, with over 55% of B2B reportedly using it for this purpose. In addition to making content recommendations, machine learning can also be used to identify prospects, personalize email campaigns, optimize paid ad spend, engage onsite, and identify broad market trends.
Takeaways
We began by saying that marketing is both science and art. The relationship is not binary. Contrary to popular belief, science can inform art. Combining technology and an analytical approach can result in improved SEO rankings, better performing PPC ads, an engaged social audience, a website that converts, and much more.
A data-driven approach helps brands get closer to their customers, move ahead of their competitors, and capture more market share. This isn’t an idealistic future of what marketing will become – it’s all possible today. The question is, are you ready to join the thousands of other marketers that are taking a data-driven approach today?