What does it mean to be a data-driven marketing success in 2015?
What does data-driven success in marketing actually look like in 2015?
I’d like to address some of the more accessible tactics that are pervasive across the top 20% of the market, not one or two outlier examples.
Micro segmentation over 1:1 personalisation
Even when data is readily available to inform highly targeted engagement, someone actually has to produce the creative and copy to trigger the engagement.
According to Gleanster, the use of personalisation in email campaigns results in 350% higher conversion over generic copy. What’s interesting about that is that 76% of organisations report they primarily use CRM data and basic demographic data for personalisation (not behaviour triggers and other implicit data).
That means the vast majority of value from a personalisation standpoint is currently derived from basic segmentation. But only 5 out of 10 marketers indicate that they are effective at segmentation. Segmenting and targeting copy and creative remains an untapped opportunity to maximise marketing spend for over half of organisations.
When it’s overwhelming to think about engaging buyers across each stage of the customer lifecycle, don’t let that bleed into engaging high-priority segments of your target audience in one stage of the customer lifecycle – by role, title, region, income, and other basic demographic attributes. Shockingly, only about 3 out of 10 organisations take the time to target segments of customers with unique messaging.
This is, of course, a lot of work for marketers, even if you cherry pick 3-5 segments. You need domain knowledge about the target audience, unique creative, copy and a refined value proposition. From a data-driven decision standpoint you should get started on prioritising your target audiences by reviewing your customer data and looking for the 3-5 largest segments of current buyers.
There are always attributes that uniquely segment your customers – when you uncover them you can isolate which areas to focus your finite marketing resources on targeting with creative, copy, and messaging. If you know what your current customers look like, use that to inform how you target new customers. That literally means reducing the volume of spend that goes toward generic brand awareness and casting the acquisition net wide looking for any opportunity.
Automating up-sell and cross-sell campaigns
Marketing is the only function in the business that actively communicates across the entire spectrum of the customer lifecycle, from the inquiry to a loyal customer. That raises two very interesting questions that data-driven marketing has answers for:
- Should marketing own the customer lifecycle?
- How should marketing allocate time, budget, and effort across the customer lifecycle?
It turns out marketers are spending disproportionally more time on customer acquisition and virtually no time on customer retention and expansion. In fact, the average B2B mid-size firm gets 70% of its revenue from customer acquisition efforts, versus the average Top Performing firm that gets 50% from up-selling and cross-selling according to a recent report by Gleanster Research and Act-On (Rethinking the Role of Marketing).
Think about it: customer data that is readily available in CRM can be used to initiate revenue-generating campaigns to customers at the later stages of the customer lifecycle. Revenue from customer engagement is not only more profitable, it can be attributed to marketing. As marketers you should be using available customer data to automatically trigger campaigns for up-selling and cross-selling to known customers within tools like marketing automation.
More importantly, marketers can finally link conversion to real revenue and demonstrate their impact on up-sell revenue. There are actually finite products and services that marketers would need to configure automated campaigns around, and for the most part these campaigns are 'set and forget' – and triggered by data in CRM. There’s no better opportunity for marketing to demonstrate their influence on revenue and use data to derive leverage from campaign creation efforts.
A/B testing on landing pages and email campaigns
According to the 2014 Gleanster Marketing Resource Management report, only 60% of small and mid-size firms conduct A/B tests on email, landing pages, and website properties. It’s actually shocking to learn how much you really don’t know about your customers when you run A/B tests on creative and copy.
This is an area that is supremely overlooked by marketers, and the capabilities often exist in multiple tools including web content management, email marketing, and marketing automation. It can, however, be a time-consuming endeavour, and many marketers are reluctant to even guess at what to test. But you can learn a ton about your target audience by running A/B tests. From a credibility standpoint, if you have aspirations to be perceived as a data-driven marketer, go test assumptions that have been made by the organisation online. Your boss can’t argue with the numbers, whether they prove your strategy is on the mark or needs some revision. This is low-hanging fruit, and it’s underutilised.
Machine learning is your best friend
One consistent theme that keeps coming up in our advisory sessions is that marketers want help in data analysis. Thanks to advances in computing power, data analysis that previously took days can now be done in seconds and often in the cloud. Machine learning applies rules to data sets and looks for correlations between data. Does this do the job of a marketer? Heck no! What machine learning does for marketing is help isolate trends that should be investigated further. Marketers still need the context about customers and products to translate those correlations in the data into action.
Today, there are predictive learning and data visualisation tools that will automatically apply rules to data sets and colour-code correlations – all these tools require is the skillset to load the data. Turnkey 'machine learning' solutions are still in their infancy, but are rapidly gaining traction as low-cost, user-friendly ways to augment analytical capabilities.
For very large data sets, machine learning is really the only way to uncover trends manual human analysis would likely overlook. The caveat here is that we tend to see predictive capabilities primarily used in mid-to-large organisations with more resources (people, money, skills) to leverage the tools. Also, the first-generation tools are priced a bit high for the average small or mid-size company. Keep in mind that the quality of the data determines the quality of the analysis, which is a huge obstacle for most organisations. However, over 50% of large enterprise firms that participated in the 2014 Digital Marketing survey had used predictive tools with tremendous success, and marketers in particular were bullish on the future potential of these tools for their organisations.
Ian Michiels is principal & CEO at Gleanster Research.