How the role of the marketing analyst is changing in the age of AI

• AI tools are enabling some mid-level analysts to deliver business solutions quickly, even when their understanding of technical fundamentals is incomplete.
• Marketing analytics has shifted from occasional large experiments to running dozens of small tests simultaneously.
• Companies are increasingly adapting hiring processes by introducing practical tasks and situational tests rather than relying solely on CVs and technical interviews.

Stay Informed

Get Industry News In Your Inbox…

Sign Up Today

Iaroslav Kobozev is head of marketing analytics at AppQuantum.

What to expect from a marketing analyst in 2026, how the market has changed, and what does AI have to do with it?

Over the past three years, the marketing analytics market (and beyond) has undergone tectonic shifts. Middle and mid+ candidates with relevant technical experience show up to interviews, and sometimes don’t understand the fundamentals.

At the same time, that lack of foundational knowledge doesn’t always stop them from solving business problems on tight deadlines, thanks to AI.

These candidates are still rare, but more are coming. We’re on the brink of a new technical era that will change everything from hiring processes to department workflows.

But before we get there, I want to share what we’re already seeing, how our expectations are shifting, and how interviews are evolving.

What an analyst should always possess

The most important traits for any analyst are critical thinking and the ability to identify pain points in the business and suggest optimal solutions. This is more important than knowing Python, SQL, or JavaScript, although without a basic technical stack, you probably won’t pass the interview.

Marketing analysts are the people who calculate returns daily. The core task of the entire department is to measure the profit from marketing activities. We invest $100,000 in an ad campaign. What’s the return?

You’re always thinking in financial terms and KPIs, building forecasts, and working closely with the finance team. It’s math-heavy work, though communication and storytelling skills are still useful.

Product analysts go deep into the product. They understand the audience, user behaviour, and how features work. Strong product analysts see the product through the eyes of the player, test that view against the data, and assess how changes affect UX. Essentially, it’s creative work. It also takes empathy.

I’ve seen people without creative intuition struggle in product analytics. And on the opposite, people without a math background who struggle in marketing analytics.

The key difference between junior, middle, and senior analysts is independence and initiative. A junior needs supervision every 2–3 hours. A middle manager can take a broader task, split it into 2–3 subtasks, deliver within a day or two, and wait for new input.

A senior needs supervision only once per sprint, week, or even two weeks. They hear a problem, think through the solution, and deliver results that directly impact the company’s bottom line.

From here on, we’ll be talking specifically about marketing analysts and teams.

Three types of skills - and why hard skills are less central

I usually group skills into three buckets: hard skills, soft skills, and business skills.

Hard skills depend heavily on the specific task stack. Typically, an analyst is either a researcher/forecaster or an engineer. A data analyst needs visualisation, automation, and manipulation skills - SQL, Python, Tableau. A forecaster needs Python, statistical methods, and data science skills. An analytical engineer needs strong SQL, orchestration tools like Airflow, and dev skills in Python or JavaScript.

Marketing analysts typically need SQL, Python, and experience with attribution tools like Upslider or Justice.

Again, hard skills map to the level. A junior knows the basics, SQL, Python, math stats, dashboards, and needs a senior to guide them. A senior, though, should already understand ML, data science, how to run parallel calculations, and optimise compute workflows.

Soft skills are about storytelling, systems thinking, and teamwork. Analytics is teamwork. Business skills are product thinking, process design, and hypothesis-driven decision-making. You also need strong metric discipline, knowing which metric to move and whether it matters.

With that on the table, let’s discuss the current state of things.

In the past three years, the importance of hard skills hasn’t vanished, but their role has changed. AI has reshaped the hiring market and heavily affected the industry altogether.

Mid-level analysts in 2020 learned their craft through dev forums like Stack Overflow, senior mentors, and long hours of reading books. Many have entered the profession during the 2000s economic boom, and they’re used to deep, extensive work and polished solutions.

Now we’re seeing a new generation entering during the neural network boom. They can get answers in seconds without understanding the underlying logic.

In the past, mid-level candidates could usually answer fundamental questions. Now? Not always, even for basics, especially without running it by AI first. Not the most perfect turn of events, but that’s the new reality we’re stepping into.

I’m from the “older school” and I still value fundamentals. But I don’t ban AI, I encourage it unless it sabotages something. As a manager, speed matters. And people who know how to use AI well can move incredibly fast.

That said, traditional companies still hire the old way exclusively: resume, technical tests. But we’re moving toward a world where someone with strong AI skills and no six-year resume can deliver senior-level results. 

Though actual experience stays highly valuable for now, hard and meta skills will still matter - it’s just that the tech market and gamedev are transforming. I’d say we’re entering an era of industrial-scale delivery and fully streamlined, assembly-line workflows.

How analytics priorities have shifted

Today, analytics isn’t about breakthroughs, it’s about making small, consistent optimisations that add up to measurable gains. One A/B test every six months has become 100 a day. Speed is everything.

Previously, analysts pulled data manually, processed it, and handed it off. Today, that’s largely a thing of the past. The role has evolved: instead of wrangling spreadsheets, analysts are building systems that let business teams access the data they need on their own, already packaged in the right format.

With automation taking over the grunt work, analysts can now focus on higher-level, strategic questions.

Building full-stack, end-to-end solutions that empower teams to plug data directly into their workflows became a normal part of the process for analysts. And the analyst is the one who handles all the context-switching, too.

Another important shift is the sheer volume of data you have to monitor nowadays. The amount of data and metrics is growing, and analysts now guide business teams on what to look at and how to interpret it. The number of experiments is growing, and so is the cost of being wrong.

You used to run two marketing experiments per quarter. Made a mistake? No worries, it could happen to anyone. Now we run a dozen experiments in parallel across multiple channels. Marketing is fundamentally a game of expansion - new audiences, new tags, new ad networks - and each move demands a fresh round of hypotheses and testing.

That’s why more and more analysts are pushing for meaningful insights, the kind that help you build larger user samples and evaluate performance accurately.

That’s why market demands are shifting. The demand for fundamental specialists with deep expertise isn’t going anywhere, in fact, it’ll only grow as the industry evolves.

But at the same time, we’ll also need an entire army of prompt engineers who don’t need that level of depth. What they can do is drive a huge number of small, fast, compounding changes. You used to spend months looking for a silver bullet, now it’s about increasing performance by 2% again and again.

All because the user payback period is much longer today. Instead of breaking even within days or weeks, it can now take more than a year. This shift is driven by market saturation, tougher competition, and ad networks claiming an increasing share of the margin.

The market is getting progressively harder to navigate. With the hyper-casual hype train going full speed a few years ago, you could develop a whole game for $1,000 and hit profitability quickly. Now, unless you have something truly special, that’s not nearly enough.

Ad creative teams need to be stronger, too. No one is going to click on whatever you put in front of them anymore just to check out a new game. Players want high-quality ads and playables that reflect real gameplay.

At the same time, marketing is moving toward data privacy. With less visibility, teams are working more blindly - more hypotheses, more experiments, more attribution models. That’s why in-house attribution is growing.

All of this has changed how I hire. I no longer rely on standard interviews. I assign practical tasks, including logic and communication tests.

When the skills test becomes a game

Here’s a recent case: I ask a candidate to explain ridge regression. They can’t. I ask about overfitting. No clue. Then I ask about multicollinearity. They answer “ridge regression helps with that.” Full-circle moment. Clearly, they were consulting with ChatGPT mid-call, scrambling for answers on the fly.

It was blatantly obvious that their knowledge was fragmented and lacked the depth to handle serious tasks. Even AI couldn’t patch that gap.

Another question I like: “Why didn’t you try a different approach?” If someone stops at “I wasn’t asked to,” it shows they wait for orders. That’s fine for juniors. But for mid-level+? Questionable.

When it comes to soft skills, the two most important qualities in an analyst are strong communication and storytelling. And by storytelling, I don’t mean just narrating a process, I mean turning complex research into clear, compelling insights and presenting those insights in a way that makes decision-making easier.

A good analyst knows how to highlight the essentials, lay out a clear thesis, and propose actionable solutions.

Let’s say there’s an early-day retention drop. A good answer: “Retention is down in early cohorts. It’s tied to these ad platforms. Within those, optimisation is down to these creatives. So we either pull back spend on these segments or pivot the creative concept.”

You can’t always gauge soft skills through a traditional interview, this is where situational prompts come in clutch. For example: You’ve finished a task, reached out to the client, and they’re not responding. What do you do? The answer usually reveals how the person handles uncertainty and how well they’ll manage communication under pressure. If they can’t follow up or push for clarity, collaboration will be tough.

Another common scenario is when someone says, “I was too shy to suggest the idea.” Analysts aren’t just data crunchers, they’re business-minded contributors. They need to be comfortable proposing new ideas and experimenting without waiting for permission.

A good litmus test: ask how they’d approach a feature implementation if there were no clear task description. Some will say, “I’d clarify the business needs myself,” while others freeze. That tells me a lot about how systematically they work.

Stress-testing questions like these don’t just help us evaluate the candidate, they help the candidate understand what kind of environment they’re stepping into.

A few more notes

Should analysts be gamers themselves? When it comes to product analytics, absolutely. A good product analyst plays regularly and might even watch others play. They need to think deeply about the product, understand player behaviour, and connect that back to the design and mechanics.

That’s not necessarily required in marketing analytics, but we still play. We love game development, and we talk about it all the time. We constantly analyse gameplay, creatives, how closely they reflect the core game, how strong the first-time user experience is, and how well it grabs attention. That kind of insight is valuable even in marketing analytics.

Team size. This depends on complexity, deadlines, budgets, and everything else we've discussed above. In some cases, 2–3 world-class specialists can cover a massive scope and drive significant business results.

In most teams I’ve worked with, the typical ratio is about one analyst per 7-8 marketing managers. It also makes sense to have a few junior interns around - but to make that work, you need a more structured approach to problem-solving, so new hires can learn from the more experienced team members.

Automation. In marketing analytics, data collection and consistency are usually automated. We pull data from a variety of sources, and both collection and processing are well-suited to automation. Analysts rely on off-the-shelf infrastructure tools and layer on in-house data collection rules.

For example, we're currently scaling our Incent Traffic program. This model rewards users for completing a certain number of levels, and we work with networks that specialise in delivering that kind of traffic. Our current focus is on infrastructure and analysis, figuring out how to measure its true cost-effectiveness.

Forecasting also fits well into automation pipelines, we try to systematise everything. Any piece of research, no matter how unique, follows a multi-stage workflow. That’s why I always tell my team: save your scripts, they’ll save you later.

In conclusion: Breaking into game analytics without relevant experience

Truth be told, there’s no guaranteed step-by-step guide, I got into game development from a completely different field myself. But in many ways, I think that “closed door” feeling newcomers often face actually helps them succeed. It forces you to get creative and find ways in.

Let’s say you don’t have relevant experience, but you really want to get into the industry. You could offer to solve a real problem for free in a month. If someone came to me with that attitude, completed the task, and did it well, I’d personally vouch for them to upper management. It’s a win-win.

It’s also great when candidates take the initiative to build something related to game development on their own - like user behaviour forecasting, churn prediction, average order value modelling, or exploring growth opportunities for UA channels.

There’s plenty of open data out there. Platforms like Kaggle and public data science competitions offer great datasets you can use to build a feature forecast.

And yet, a surprising number of CVs don’t even include the basics, like a cover letter. From a hiring manager’s perspective, that letter is essential. It tells me what drives a person, how well we might click, and whether we’d work comfortably together.

One more tip that often gets overlooked: keep your portfolio up to date. Document your projects, results, blog, and social media posts. If you talk publicly about what you’re working on or learning, it stands out - because almost no one else does it.