After a decade of maintaining the status quo, the marketing analytics field has experienced significant changes to measurement and reporting on an almost annual basis. These changes have brought new challenges to marketers forcing them to adjust strategy, tactics and budgets. The seminal moment that kick started this change was Apple’s hard stance on data privacy and uprooting not only how iOS marketing is measured but also serving as a dagger to the key revenue streams that have lifted publishers to the top of the technology industry. In only two years since this iOS event occurred, Android and popular web browsers have made similar changes impacting digital businesses regardless of the platforms they sell their products on. The simple days of growth at all-cost business models and access to an abundance of data to optimize around are gone and with it, constant change and the need for businesses to be nimble and smart. In 2023, I anticipate this volatility to continue with three key trends from the past few years playing a major role in my predictions and shaping the marketing analytics landscape in 2023.
Three Trends Shaping 2023
User Data Privacy
The previously mentioned iOS event makes it evident that businesses need to change their approach to marketing measurement as deterministic third-party tracking via pixels, fingerprints, and SDKs fade away. While businesses running on iOS have already needed to adapt to an almost complete removal of third-party capabilities, I have surprisingly found most businesses have lacked a measurement strategy that adapts to the reporting limitations presented. I’m not sure why this might be, but given the current economic conditions, marketing budgets will be under the microscope with a top priority for all businesses to have a proper measurement system to guide growth teams in the right direction.
There is no argument that technology has been one of the hardest hit industries by the economic climate since the second half of 2022. More than a decade of almost zero interest rates drove high valuation multiples and a growth-over-profit mindset for most technology businesses. Investment flooded into public and private markets making technology one of the top-performing industries. However, all of this was flipped on its head after various macroeconomic conditions gave way to rapid inflation and government policies alleviating these pressures. Most of the policy changes have been raising and maintaining interest rates at their highest in decades. With less lending due to interest rates came a reckoning of the technology markets that we are seeing play out today. In its wake, there have been mass layoffs, startups going defunct, and a systematic shift in businesses focusing on positive cash flows and profitability.
AI took the industry by storm in the last half of the year and filled the tech buzz void last felt with the cryptocurrency rush a few years ago. Despite the field entering the “toy” phase on the user enthusiasm scale, there have been substantial breakthroughs and bets in the space that will continue to make it a dominant topic of discussion in 2023. In the marketing space, its immediate influence has been in the creative field. Large Language Models and Generative AI have been able to produce copy and visual assets with a simple phrase in a text prompt.
Placing My Bets
These three trends play a significant role in my five predictions that follow. While I don’t anticipate getting every prediction correct, my hope is that this will bring awareness to the evolution playing out in marketing and emphasize the importance of having marketing analytics resources to help measure and guide your business’ growth.
Consolidation of the MarTech Space
Last year we saw a dramatic drop in the public market performance of technology companies. The NASDAQ dropped 33% since the start of the year with high-interest rates wiping out valuation multiples and with it investor appetites in not only public but also private markets. With less lending and budgets cutbacks, massive layoffs across both technology goliaths and scrappy startups have sadly become almost a daily occurrence. I, unfortunately, don’t anticipate the negative news to stop in at least the first half of 2023 with so many economists forecasting a recession (albeit a mild one) to ride out until 2024.
What this means is that we won't likely see the bleeding stop and start to see more bankrupt companies as revenue takes hits, funding dries up, and investors balk at anything less than certain. Given the exponential growth in MarTech tools in recent years and a personal sentiment that most of the businesses in the space are features rather than tools, I think that this is a space that will see a significant amount of contraction as the year carries on.
In a typical technology business, IT/Engineering tends to see the largest budgets due to their critical importance to the success of the business. There is less scrutiny over the allocation of those dollars and it is tougher to find areas for reduction that won’t also impact revenue. In comparison, almost every department tends to have to make strong cases for tools. While there will always be one or two tools that are integral for a department to operate, most tools tend to be "nice to have" versus "must have" and given the current market, it is very likely that there will be budget reductions that will put most tools on the chopping block. What this means to marketing departments is that every MarTech tool will be under the microscope.
In my decade of experience in marketing, I have found a lot of difficulties finding real value in most MarTech tools I have come across. Their product is usually built to solve one subset of issues in otherwise large and complete applications that are commonly used. These products are less focused on the important value propositions of driving revenue or saving money, but rather cutting down on resources needed. While this is always important because “time is money”, the cost savings tend to be significantly less than the cost to the business by paying for the application. I’m not alone in my sentiment towards MarTech tools as Casey Winters points out other cases that make it difficult to receive full buy-in from their audiences. This lack of “must have” appeal will likely lead to most tools being cut from the budget, but without much of an impact on the businesses that use them.
Increase in Data Unification Adoption
It is surprising to me that this is still something that would be talked about today given the advancement in data integration tools and the ubiquity of reporting API’s for important marketing platforms. However, I have found that an overwhelming amount of companies that I have worked with still do not have their data unified and automated. There is certainly no shortage of tools or resources like plug-and-play tools, FiveTran, and Airbyte, which have eliminated the heavy burden of understanding and using various marketing platform APIs, or spreadsheets plugins like Supermetrics, automating data into the practical environment of a spreadsheet. There are endless ways to unify and automate your marketing data pipelines, but most businesses are still relying on the mundane process of exporting a report to CSV file, formatting the data, and then copying and pasting it into another spreadsheet with a custom schema.
Maybe it had been the luxury of a bull market where you could afford to have dedicated roles for these tasks. But given today’s market conditions and the pressure on marketing programs to deliver with less than before, the key to success for a program is speed and efficiency. Reporting automation delivers both of these traits and I believe that the value created with data unification tools will finally be realized this year and lead the last holdouts to finally adopt these tools into their marketing analytics infrastructure.
Incrementality Goes Mainstream
This prediction has technically already been in motion ever since recent digital privacy changes created measurement woes and forced many to revert back to traditional marketing measurement. However, just like data unification, the adoption of these has been slow. Incrementality has traditionally been an afterthought in how most businesses think about measuring marketing performance and strategy not only within marketing but the business as a whole.
In many cases, the hesitation to adopt this methodology has been due to two common reasons. The first is the requirement of statistical modeling which requires certain skill sets that can be costly and resource intensive. The other is skepticism around the results due to familiarity bias. We have based our decision-making on the trust of analytical tools despite known limitations in tracking and telling only part of the attribution story. We do this by nature because it is easier to trust well-known names behind the technologies and the comfort of how reporting is universally accepted, rather than challenge the integrity of what these platforms say and be experiment minded to prove the true impact of the program.
The trust you can have in these platforms as a source of truth for reporting is rapidly decreasing, with tools like marketing (media) mix modeling and experimentation paving the way as future-proof approaches to measuring the impact that each marketing effort is having on the bottom line. What this means is that there will likely be an uptick in data science demand to set up the reporting infrastructure of the future and build out these models. New products will also make these techniques more approachable and accessible to marketers. The latter has already started to play out already with tech giants open-sourcing libraries to streamline the adoption of incrementality as a framework to make sure marketing programs are running as efficiently as possible. We will only see more contributions and advancements in this space over the course of the year due to more privacy enforcement and the scrutinization of marketing performance.
The True Value of “AI” is Still Too Early to Determine
Artificial Intelligence, the trendy technology topic of 2022, exploded onto the scene with fascinating results produced by generative models. Generative AI and Large Language Models (LLM) (e.g. ChatGPT) are two of the most popular AI tools at the moment, creating impressive results with only a few words or sentences written into a text prompt. While your imagination is one of the few limitations of using the tools, the results these products have produced for a business have not been that strong and likely keep the verdict on their value out until 2024 at the least.
LLM has already started to be used by companies for anything from marketing copy to sales pitches, but despite being helpful in uses of summarization and inspiration, it takes a good amount of prompt engineering to arrive at satisfactory results and isn’t guaranteed to be accurate, as many have pointed out in criticism on social media websites. Building on this criticism, I have also felt like personality and punch are missing from the LLM results which are key to any marketing material. It is what captures the audience's attention and sells them in the short amount of time you have their attention. These are traits that make copywriters so important in the context of marketing and advertising and something that hasn’t been reproduced in the work I have seen. While not there yet, I do believe LLM is the furthest along in its applicability to everyday needs in marketing.
Generative AI on the other hand seems less viable in my eyes due to model preferences in generating images with artistic styles over the graphical design styles most commonly used for marketing purposes. Most Generative AI examples you observe on the internet tend to be in an illustrated or artistic style that in some cases can be mistakenly taken for original works. However, you would be hard-pressed to find an example of an image that replicates the necessary structure of an advertisement. They lack the necessary branding, still struggle with producing certain elements like human hands, and are rooted in artistic composition, rather than the layout and composition found in graphic design. For these reasons, I believe that Generative AI is further off in its business use cases than LLM, but I could imagine a day when a model is built off of a graphic design training set that could deliver results that are a step removed from being production ready.
Focus on 1st party data and segmentation
As 3rd party data’s influence on marketing starts to diminish, the focus will be on optimizing around 1st party data and with an importance on the ability to create audience segments that are reliable and most impactful to the bottom line. This should already be a focus for most businesses, but rarely a top priority for marketing as a stakeholder for growth. Most of the focus for marketing has historically been to use pixels to help optimize a platform's algorithm (ad, social media, etc.) and find your users. With those platforms also impacted by privacy changes, there is less that we are in control of tactically and more time that will likely be dedicated to utilizing 1st party insights and segmentation testing to deliver results to maintain the growth.
Stay on Your Toes This Year
There is plenty of uncertainty and change that has been playing out quickly as the new year gets underway. With most businesses bunkering up for a year of volatility, marketing strategies and budgets will change frequently over time. Profitability and cash-flow positive are the new names of the game. Having a strong marketing analytics program in place will be important to accurately understand the effectiveness of the program, make smart bets on what the business needs and increase the likelihood of making the right strategic decisions.