Surprisingly, though, Gartner’s fresh CMO Spend Survey indicates that budget allocation for martech reduced to 26% this year.
To some extent, this not-so-drastic plunge can be attributed to technology purchase cycles. Considering that investment rates peaked in 2017-2018, we can extrapolate that now most businesses are deep into exploring what they have purchased and how the new tech can be used to optimise in-house marketing processes.
According to Gartner, so far marketers have leveraged only 61% of the functions available in their martech portfolio. The reason? Most organisations lack people, skills and a proper vision for their marketing analytics programme.
Ascend2 research further mirrors that sentiment indicating that 61% of companies originally invested in martech solutions to improve marketing efficiency and marketing ROI. Yet, post-adoption, 47% of respondents still struggle to capture that elusive ROI figure and 40% are far from being as productive as they’d like to be.
So why is it that so many are struggling with their marketing analytics programmes? Let’s take a look at four of the most probable reasons.
1. Lack of well-articulated, measurable use cases
Too often, marketers get overly enthusiastic and think that a single analytics solution can transform every aspect of their digital marketing operations. As a result, they approach the selection process by comparing a tool against a list of ‘wants’ and desired future benefits – reduced ad waste, better attribution modelling, content analytics and so on. But their list lacks clear, KPI-backed use cases, specific to their organisation.
This is problematic because you certainly can purchase or develop a content recommendation algorithm similar to Netflix but doing so won’t necessarily increase your revenue. This might be because:
- Your IT infrastructure is weaker.
- Your data culture is less mature.
- You’re not even in the business of online streaming.
A poorly defined use case paired with a cutting-edge solution will equal low-to-no returns. To determine which data analytics use cases are worth pursuing, ask yourself the following questions:
- Do we have sufficient, accessible and high-quality data for this?
- Will we need new people on the team to support this use case?
- How will our internal processes need to change?
- Can we measure and attribute value post-adoption?
Ideally, you should test several options against these questions and round up to 2-3 feasible use cases, along with a set of measures you’ll use to evaluate their performance within one year. Doing so will help you prioritise your investments and scale the programme effectively.
2. Gaps in analytics and data science roles
Analytics and martech solutions bring the most value when they are overseen by skilled staff. But most organisations centre their marketing investment around tech, not people. As a result, as much as 83% of agencies and 68% of advertisers say they need more people with “data skills”.
Empowering your team with new tools that will help them accomplish more is just one piece of the puzzle. Most software will be of little use without proper training and expertise. So if your recent investment isn’t bringing the expected results, you may want to assess your teams’ digital skills and invest in upskilling.
But before you rush with any decisions, make sure that you understand how data science and marketing correlate, what type of skill sets you lack on both ends and how you can fill those gaps – by hiring external consultants, bringing in new in-house specialists or investing in employee training.
3. Over-obsession with data
Marketers often think that they should plug in as much data as possible to get the best outcomes. However, that’s not always the case. Data consolidation and cleansing is an important first step for enabling powerful analytics, but you don’t need every bit of data to gain accurate results. In fact, according to McKinsey, the ‘over-doers’ waste as much as 70% of their data cleansing efforts and end up with data lakes that are not fit for purpose half of the time.
Your end goal (a measurable, feasible use case) can require just a few data sources. Additional ones can be plugged later on when you are ready to tackle more complex problems. Furthermore, new-gen machine learning models such as those powered by deep or reinforcement learning can be trained on limited data sets.
4. Stripping analytics out of context
Customer behaviour has become increasingly complex, spanning over a multitude of digital and physical touchpoints. Learning to identify, capture and attribute those patterns on the correct business outcomes is crucial to increasing analytics accuracy.
However, numbers alone will not tell you the full story. Data scientists cannot program viable algorithms when their personal view of the customer and the bigger business picture is lacking. They may pass on an important trend as an anomaly or arrive at the wrong conclusions by over-prioritising non-essential trends.
To avoid such a scenario, ensure that your analysts:
- Understand the nature of the analysed market – trends, sales cycles, common internal and external drivers, typical buyers’ journeys, etc.
- Know the desired outcomes such as purchase funnel metrics, average conversion rate or customer lifetime value.
- Can gauge the interplay between different factors. For instance, product design changes can affect sales numbers, but sales numbers can also drop due to seasonality and not necessarily new packaging.
Getting things straight with analytics is essential to make smarter investments and scale your programme sustainably. Without articulating measurable use cases, determining the right data sources and developing the right skill sets and analytics practices within your team, even the most robust analytics solutions will bring little value to your marketing.