Editors’ note: Today’s post is by Mitja-Alexander Linss and Lou Peck. Mitja-Alexander is Senior Director, Publications Marketing at Optica, and Lou is the Chief Executive Officer and Founder of The International Bunch. Reviewer credit to Chef Stephanie Lovegrove Hansen. The audio file linked below provides a spoken version of this blog post.

In academic publishing, few things spark as much discussion as the question of what actually drove an increase in submissions. Was it a campaign, commissioning, market shifts, or community engagement? Or was it more holistic, driven forward by contributions from several teams across an organization, not just the usual suspects? Usually, it is a combination of factors, which is exactly why attribution matters.

Just because influence is hard to isolate does not mean it is not there. Beyond guesswork, the solution is to clarify what actually contributed to the increase, why, and what that means for future plans and strategies.

This is no small feat.

Attribution is already difficult in most sectors, with long journeys, multiple touch points, disconnected systems, and competing models of first touch, last touch, and everything in between, making even well-resourced teams struggle. Add in changing discovery habits, agentic Artificial Intelligence (AI), and zero-click behavior, and the path becomes even harder to follow. Scholarly publishing just adds another layer of complexity.

Attribution is not simply about who clicked on what. It is about understanding what may have contributed to a manuscript submission, and in some cases, whether that submission eventually became published content. To design outreach that best supports both journals and authors, we need to make sense of a journey that cuts across content platforms, search, campaigns, calls for papers, conferences, editors, societies, submission systems, author management platforms, institutional relationships, etc. It is rarely linear and almost never housed in one place.

So, How Can Publishers Measure Attribution Effectively?

1. Ask the researcher

Capturing why someone submitted to a journal in the submission process can go a long way. This can be in one of several touchpoints, including your submission platform, a survey after submission, or a pop-up during key parts of the production process to sense-check satisfaction.

Self-attribution is not perfect, but it is practical, quickly creates a useful baseline, and provides a direct source of insight that is too often missing. Make it easy, keep the options relevant, and ensure the language is clear. Include an “other” field so patterns can emerge rather than being forced into the wrong box.

Self-reported data does come with caveats, so treat it as one layer of understanding, just not the whole picture. Distinction matters as attribution debates become trapped between two extremes: either a data point is treated as definitive proof, or it is dismissed entirely. In practice, we need to get more comfortable working with directional insight, provided it is used carefully and in context.

2. Make the journey easier to measure

If you want better attribution, you need a better view of the customer journey itself, as this is often where things break down. How did the author find you, and how do they get to achieve their goal? Every disconnected system creates another blind spot, making the measurement layer even more important. Systems that offer a unified view of engagement activity give teams a huge leg up in identifying attribution.

Dashboards, rankings, or internal scoring models can also be part of the solution for quantifying marketing influence. This structure should be co-created with key stakeholders to ensure better buy-in. The moment a proxy metric is presented as certainty and without key input, skepticism grows, and it loses all merit.

Build foundations with consistent tagging, agree on campaign naming conventions, better event setup, improved interoperability between systems, and shared definitions of what counts as engagement, progression, and conversion. Take Google Analytics as an example: set it up properly, define key events correctly, and measure sessions seamlessly across different systems or apps used on your website. This is what turns reporting from opinion into something far more trusted. We need a measurement approach that is disciplined enough to reduce the number of assumptions.

3. Understand who they are and what they need

Mapping your researcher persona means recognizing who they really are, what they genuinely need, and how you can help, which is essential if you want to stay relevant. It also means understanding what they need in their own region, where the geopolitical landscape, accessibility requirements, and expectations can all look very different. Remember too that researchers wear many hats. Across a single career, someone might be an author, a reader, a peer reviewer, an editorial or advisory board member, a society member, a volunteer, an advocate and ambassador, and one day maybe even a future staff member.

This is also why we are moving from omnichannel marketing to optichannel marketing. We used to operate in a multichannel world that was all about quantity, being everywhere and always available, but that approach created friction and demanded huge resources. Omnichannel shifted the focus toward seamless integration and building everything around the customer. You will sometimes hear this described as the “North Star.” It is worth noting that the customer does not always require a financial transaction; even a knowledge exchange is a valuable transaction. The trouble is that channels evolve faster than anyone can integrate them, so you can feel behind before you have even finished.

Optichannel moves the focus from integrating every channel to optimizing the ones that actually matter. Which channels are right for your community, your geography, your journal, your subject? Sometimes people webroom, browsing online and doing their research before they engage. Being deliberate about where you invest is essential – not everywhere, just where it counts. This aligns with moving into a post-digital world where physical, digital, and virtual customer journeys all combine. We need to nurture our communities from early career through to retirement, which means understanding the different personas, pain points, cultures, and challenges along the way. Localization becomes imperative.

4. Fix the data foundation

Attribution gets so much harder when the data beneath it is fragmented, incomplete, or difficult to use, which is more common across our industry than you think.

If you are lucky enough to have a connected data lake, you’ll have a better story to tell, or a martech solution that helps you make more informed decisions with better segmentation and targeting. But what if your author data lives in one place, institutional data in another, campaign data somewhere else, and journal and book performance data somewhere else again? Sound familiar? Teams then try to answer strategic questions from a patchwork of systems and spreadsheets, and wonder why confidence in the answers is low.

Getting the data foundations right does not require an overnight transformation project. Are author records captured consistently in your Customer Relationship Management system (CRM) and linked in your Customer Data Platform (CDP)? Can you segment by journal or book, institution, region, discipline, or topic without relying on multiple teams each time? Can contacts and accounts be linked to pipeline and opportunities in a practical, usable way? These are not small operational details; they determine whether attribution is even possible.

Better data does more than improve reporting. It helps marketing teams run more relevant campaigns, spot patterns across publications and subject areas, and decide where to invest with better insights. That may lead to a CDP, a data warehouse, or a more connected data environment, or it may simply mean far more discipline around the systems already in place. Whichever way you go, the principle is still the same: fragmented data limits attribution. Better data makes more informed decisions possible.

5. Understand behavior for better choices and segmentation

One of the biggest missed opportunities is failing to act on the signals already there. Researchers tell us a great deal through their behavior. These are meaningful indicators of interest, and when we respond to those signals with relevant, timely messaging, performance tends to improve.

This means delivering relevant content that offers value and resonates on a human level — especially when you are interacting with humans and the AI tools and bots they instruct. Research communities are busy, selective, and used to better quality digital experiences elsewhere. Generic messages are easy to ignore, especially in a world where mass personalization is omnipresent. Contextual messaging and advertising are more likely to land.

There are industry-specific examples of this kind of capability emerging. One is Hum, a behavioral audience engagement and customer data platform. It helps align with where the market is heading, with a better connected understanding of anonymous and known user behavior, stronger audience intelligence, and a direct connection between engagement and publishing outcomes. As this use case shows, on average, conversion rates of around 2% are three times higher than those of traditional display advertising. Over 18 months, Hum conversions directly influenced more than 1,000 journal submissions and associated publishing revenues, with 73% submitting to multiple journals and indicating sustained engagement.

6. Support cross-functional teams

Strong attribution comes from working better with marketing, editorial, product, community engagement, production, peer review, data/analytics, IT, sales, commercial, and operations colleagues, too. By bringing these perspectives together early on, the publication marketing strategy becomes more relevant, credible, and effective. It goes beyond just tactical output, and that matters for attribution too.

So, if editorial and marketing shape annual priorities together, plan calls for papers together, and agree on shared goals from the outset, it becomes much easier to understand what may be contributing to interest and where momentum is building. It also makes it easier to prioritize what is most important when we are resource-strapped. It reduces one of our biggest barriers: siloed work that slows decision-making, weakens campaigns, and creates mistrust before results are even reviewed.

This is where the conversation needs to mature. Marketing is just one part of the picture, just like editorial direction, technology, accessibility, reputation, timing, community standing, and subject interest. Each matters, but the fact that impact cannot be proven perfectly does not mean it is not there; better attribution helps us measure contribution more clearly.

Good Attribution is not Perfect

Publication marketing will always involve complexity, multiple systems, and variables that no team can fully control, no matter how well resourced you are. Marketing may influence the route to submission, but it should not be expected to determine the result. That belongs to the publishing process. The real need is not a perfect model but a clearer, more credible way to understand contributions. We are also an industry too focused on measuring quantity, and not quality. When we deep-dive or reverse-engineer a question, we can find compelling information we didn’t even know existed.

By creating a more collaborative, joined-up view that shows where activity is working, where momentum is growing, where data is weak, and how teams can improve over time, we can build better measurement and improve our outcomes. In the end, attribution isn’t about what a single activity did. It’s about how we understand the customer and publication journeys and life-cycles to make more confident, informed, and effective decisions, and about incorporating more ways to ask those who matter: the researchers. So start with the question you need to answer, then work backwards to find the right way to measure it, whether the answer is what you expect or not.

AI Disclosure: Google Gemini was used during the initial research phase to analyze posts in The Scholarly Kitchen, scope the article’s focus on attribution, and verify adherence to house editing guidelines. Initial editing and grammatical checks of the drafted text were performed using Gemini and Grammarly. All AI outputs and editing suggestions were subjected to full human review and final verification by the authors.