The challenge for marketers has quickly shifted from not having enough data to make informed decisions to having so much data that it becomes nearly impossible to organize, present, and act on it in any meaningful way.
I had the opportunity to talk to Tara Robertson, Chief Marketing Officer at Bitly, to get her insight into how modern businesses can better understand what data really matters, how to collect it, and how to move away from making 'gut-feel' decisions that could be costing your business.
How is AI changing how potential customers interact with marketing content?
The clicks that do happen mean so much more.
Tara Robertson, Chief Marketing Officer at Bitly
AI has changed the early stages of the customer journey pretty significantly. People still want to make informed purchase decisions, but increasingly, they’re researching and comparing options inside AI tools and LLMs rather than exclusively clicking through a traditional search results page. That creates a much bigger “dark funnel” for marketers because more of the customer journey occurs in places we can’t easily see or measure.
At the same time, the visibility signals we’ve historically relied on are becoming less reliable. Ranking highly on a SERP, for example, doesn’t necessarily translate to the same level of traffic when AI Overviews can give someone the information they need without requiring a click.
On the other hand, since there are now fewer clicks than before AI became so intertwined in discovery, the clicks that do happen mean so much more. If someone looks at the AI summary, gets the answer they need, and still clicks, that shows real intent, not just casual browsing.
So, marketers should really strive to change their mindset from “how can I get more clicks” to “how do I make sure I know what's happening with the clicks I do get, and how do I create content good enough that a human or an AI system wants to surface it in the first place.
Your recent study found that 56% of marketers rely on gut instinct to make marketing decisions. With so much data now at marketers' fingertips, why do you think this is still the case?
This stat has stayed with me the most since we gathered the data. It’s not that marketers don’t want to be data-driven, but that the data they collect is spread across too many tools, making it difficult to make strategic decisions in real-time.
Our research revealed that teams are juggling an average of six different tools to measure performance, and more than a third are using seven or more. With this amount of data on hand, many would think that this provides clear campaign performance insights. Instead, the data is fragmented and typically focuses on only one channel. Very few are integrated with other tools, so teams are looking at pieces instead of the full picture.
So when faced with a tool and data overload, marketers lean on what has worked for them in the past, which is often going with their gut. This isn’t a bad strategy when the data fails to bring clear insights because those efforts worked for a reason, but in today’s landscape, that’s not what’s going to create success. Just because one tactic worked in a previous campaign doesn’t mean it will work again in the current one. My take is that the fix isn't "try harder to be data-driven"; it's giving people faster, cleaner signals that are easier to access so instinct becomes the backup plan instead of the default.
Your report also shows that 73% of marketers only realize a campaign is underperforming after it’s too late. What processes and tools can marketers put in place to ensure data is timely and helpful?
Locked-in budgets were the number one thing marketers told us limits their ability to act on what they're seeing.
Tara Robertson, Chief Marketing Officer at Bitly
My biggest tip is to be proactive.
Instead of looking at the data at a campaign level and analyzing it after it's complete, look at the interaction level and get granular with your insights from clicks, scans, and visits.
The overarching issue is that 86% of marketers wait until a project is complete to analyze data at the broader campaign level. By the time final reporting comes around, the budget is spent, and the window to optimize is closed. Analyzing granular interactions in the early days of a campaign gives you the agility to adjust in real time.
Other tips I usually share are to set a check-in point when campaigns are being built instead of waiting until after to see if it works. Set multiple, if needed, especially when just starting out.
Second, use AI to help with data analysis to accelerate the time-to-insight process. It’s still a good idea to always double-check your work, since AI can still make mistakes, but we found that marketers who incorporate AI into their workflows report receiving faster data insights and are more confident in their decisions, yet only 16% are actually doing this today.
Finally, build a little flexibility into your budget and your plan. Locked-in budgets were the number one thing marketers told us limits their ability to act on what they're seeing — you can have perfect visibility and still be stuck if there's no room to move.
How should small businesses use AI to better understand their data and make more informed marketing decisions?
Small businesses don't usually have a dedicated marketing team to thoroughly review, analyze, and act on the data that’s collected behind marketing activity. AI-enabled tools can help close that gap and surface insights to help direct the "what's next?" decision.
The goal isn't to get every new tool that comes to market, but to make sure the tools they already use provide clear insights quickly. Instead of exporting a spreadsheet and trying to manually figure out why last week's numbers moved, you should be able to just ask the tool.
I would start small. For example, business owners could ask AI assistants to provide weekly summaries of what changed. This way, if some of the numbers aren’t adding up, the issue can be detected sooner. When you are the one running the show, any way to save time on routine tasks is helpful. The businesses getting real value aren't the ones with the most sophisticated AI setup; they're the ones who made asking questions and getting answers often and quickly a habit.
Around 18% of marketers said they struggled to connect marketing efforts with customer retention. Beyond repeat purchases, which metrics can marketers monitor to better understand customer lifetime value?
Repeat purchases tell you a customer came back. They don't tell you why, or how close that customer might be to leaving anyway. To actually understand lifetime value, I look at three things beyond the purchase itself:
Customer research and direct feedback. NPS and CSAT are useful, but only if you're measuring them at the moments that matter, not once a year in a company-wide survey. Score it right after onboarding, right after a support ticket, right before renewal. The number tells you something moved; the open-ended verbatims tell you why. That qualitative layer is what most retention dashboards are missing.
Activation and usage depth, especially if you're product-led. Time to first value, breadth of feature adoption, whether usage is expanding or flatlining. A customer who renews but uses less of your product every month isn't retained; they're on their way out and still paying you for now. Watching usage trendlines catches that months before the churn shows up in your revenue numbers.
What happens in the white space between purchases. Are they opening your emails, engaging with your content, actually using your loyalty program? Rising support ticket volume or a drop-off in engagement is usually the earliest signal you'll get that something's wrong, well before it shows up as a lost customer.
None of this replaces repeat purchase rate as a metric. It just gives you the why behind the number, and enough runway to act on it before the customer is already gone.
Top-level metrics often give an idea of the way a campaign is going, but don’t always give the full picture. What micro-interactions and signals should people be looking for?
These analytics are what I think marketers aren’t focusing on enough. When reviewing success at the program and campaign levels, reporting gives you an overall view of how performance went, but it won’t provide the specifics.
More than half of the marketers who participated in our study said they focus on high-level strategy to inform decisions, while only 14% actually track micro-interactions such as individual clicks, scans, or page visits.
The small, seemingly inconsequential data points are often where the real story lies. Knowing which specific link or QR code is gaining the most traction, or even which channel it originated from. Whether a spike in engagement on one post actually correlates with anything downstream or is just noise. Knowing what happened before they converted could make the difference between a successful and failing campaign.
Even going as small as the time and location patterns can be helpful. Are most of the interactions coming from a QR code located in the store or one shared digitally? These small indicators reveal insights into customer intent. None of these metrics typically end up on top-line dashboards or reports, but when viewed as a whole, you can review which touchpoints are performing well and which ones aren’t.
How can businesses cut bloat when picking which marketing and analytical tools are right for them?
Fewer, better-connected tools will always beat more tools that don't talk to each other.
Tara Robertson, Chief Marketing Officer at Bitly
I suggest beginning with a thorough review of the tools you currently use and what you would be losing if you let it go. Tools are often added to the stack to solve one problem at a time, and a year later, those issues might no longer be relevant. Be mindful of what each offers and whether you truly need them, and always remember that you should be looking at your strategy BEFORE your tooling. Not the other way around.
Before adding a new tool to the stack, ask yourself what is unique about this product that nothing else you currently use provides. If features overlap, you don’t need to have both. I’d also consider whether those tools integrate well with the other programs you’re using. They are often forgotten about if they aren’t directly in your workflow. Fewer, better-connected tools will always beat more tools that don't talk to each other — that's basically the key finding of our research in one sentence.