If you've ever tried to get embedded analytics onto your product roadmap, you'll know it rarely goes smoothly. The business case is compelling, users have been requesting it, and leadership is usually on board. But somewhere between “let's do this” and “it's live,” things get messy.

Our Embedded Analytics Opportunity 2026 report surveyed product professionals across industries about the state of embedded analytics – and one of the clearest findings is that the barriers to shipping it are remarkably consistent, regardless of where a team sits on the adoption curve.

Whether you've already shipped, you're mid-build, or you're still weighing up whether to start, the same challenges keep coming up. Here's what's getting in the way, why these blockers are so persistent, and what product teams can do to navigate them.

Technical complexity is the barrier that won't go away

Let's start with the headline finding: technical complexity and engineering resources are far and away the biggest obstacle to embedded analytics. Among teams that have already implemented the capability, 57.1% say this was their biggest challenge. Among teams still considering it, 42.9% flag it as their top concern.

That consistency is striking. It's not like the technical challenge fades once you've shipped – if anything, teams who've been through the process are more likely to identify it as the primary hurdle.

This isn't a new problem either. In the 2022 edition of this report, 20% of respondents cited a limited number of developers as a key constraint. While the way we framed the question has evolved, the underlying story hasn't changed: engineering capacity remains one of the biggest bottlenecks for product teams building embedded analytics.

And it makes sense when you think about what's actually involved. Embedded analytics isn't just a feature you bolt on. It typically requires backend data modeling, query infrastructure, visualization layers, performance optimization, and ongoing maintenance as data volumes grow and user expectations shift. 

That's a serious engineering commitment, and it competes directly with every other initiative on your roadmap.

Data security and integration: the foundational hurdles

Technical complexity grabs the spotlight, but the next tier of challenges is arguably just as difficult to navigate – and more likely to be underestimated at the planning stage.

Among teams that have implemented embedded analytics, 42.9% cite data security and privacy concerns as a major challenge, while 35.7% point to the difficulty of integrating analytics with their existing tech stack. For teams still considering the capability, these same issues show up at 21.4% each.

These are foundational problems, not strategic ones. Teams aren't struggling with “should we build this?” – they're struggling with “how do we make this work safely and reliably within the systems we already have?”

Security is particularly thorny in the embedded analytics context. When you're surfacing data directly within a product, you're potentially exposing it to a wider audience than it was originally designed for. Role-based access controls, data masking, multi-tenancy, and compliance with evolving privacy regulations all add layers of complexity that product teams don't always anticipate upfront.

Integration, meanwhile, is the kind of challenge that sounds manageable in theory and becomes painful in practice. Most products aren't built on a single, clean data architecture. They're stitched together from multiple systems, APIs, and data stores – and getting analytics to pull from all of those sources reliably is difficult.

The challenges don't look that different at any stage of adoption

One of the more interesting patterns in the data is how similar the barriers look across different stages of adoption. Whether a team has shipped, is building, or is still evaluating, the same concerns dominate.

For teams that have implemented embedded analytics, the top three challenges are technical complexity (57.1%), data security (42.9%), and tech stack integration (35.7%). For teams still considering it, the top concerns are technical complexity (42.9%), infrastructure costs (28.6%), and a three-way tie between performance, security, and integration (all 21.4%).

The ordering shifts slightly, and cost becomes more prominent for teams who haven't started yet – which makes sense, since it's easier to worry about cost before you've committed resources. But the overall picture is consistent. The problems teams worry about before they start are largely the same problems teams deal with after they've shipped.

This has an important implication for product leaders: these aren't problems you solve once and move past. They're ongoing constraints that need to be managed throughout the lifecycle of your analytics capability. If your planning assumes the hard part is getting to launch, you're likely underestimating the effort required to maintain, secure, and evolve the capability over time.

What's holding back non-adopters?

It's also worth looking at the 19.2% of respondents who aren't currently pursuing embedded analytics at all. Their reasons paint a slightly different picture.

The most common reason is straightforward: 40% say it's simply not relevant to their product type. Fair enough – not every product needs embedded analytics.

But for the rest, the barriers are more situational than fundamental. Equal shares (20% each) say they haven't seriously evaluated it yet, it's not a priority compared to other product initiatives, or they have technical complexity concerns. This suggests that for many of these teams, embedded analytics isn't off the table permanently – it's deferred.

And the data on what would change their minds supports this. When asked what would make them reconsider, 60% say direct customer requests would be the biggest trigger. Another 40% point to evidence that competitors are offering it, and 40% say case studies showing improved retention would be persuasive. Only 20% say nothing would change their view.

For product leaders at these organizations, the takeaway is worth noting: the demand signal that tips the balance is most likely to come from your customers, not from an internal business case. 

Keeping an ear to the ground on what your users are asking for – and what your competitors are shipping – may matter more than a theoretical ROI spreadsheet when it comes to getting buy-in.

The structural problems underneath the surface

Beyond what the survey data shows, the expert commentary in our report reveals a deeper set of structural challenges that are harder to capture in a multiple-choice question.

Shubhojeet Sarkar, Senior Group Product Manager at Meta, described the biggest blockers as fundamentally structural. 

On data quality, he noted that “most teams are working with schemas designed for storage, not inference; ambiguous fields; inconsistent coverage; and gaps in event pipelines. That work has to get done before intent-aware surfaces or NL query layers are viable, and it's consistently deprioritized.”

In other words, many teams are trying to build sophisticated analytics experiences on top of data foundations that weren't designed for it. The schema cleanup, the pipeline work, the field-level documentation – none of it is exciting, but all of it is necessary. And because it's rarely visible to stakeholders, it's perpetually at the bottom of the priority list.

Shubhojeet also highlighted what he called the “org structure” problem: “The team building the analytics surface rarely owns the workflow it's embedded in. That means they can't instrument outcomes, can't run holdouts, and can't define what a good decision looks like in context.”

This is a subtler but equally important barrier. Embedded analytics is inherently cross-functional – it sits at the intersection of product, data, engineering, and often the customer-facing teams that own the workflows users are operating within. When ownership is fragmented, it becomes harder to build cohesive experiences and even harder to measure whether they're working.

How to navigate the barriers

None of these challenges has an easy fix. But the data and expert insights from our research do point toward some practical approaches that can help product teams move forward more effectively.

Right-size your initial scope

One of the consistent traps teams fall into is trying to build a comprehensive analytics experience all at once. Every dashboard, every data source, every user persona – all in the first release. This is a recipe for the exact technical complexity and integration challenges the data highlights.

A more practical approach is to start narrow. Pick a single user workflow where analytics can add clear value, build for that use case first, and expand from there. This doesn't just reduce technical complexity – it also makes it easier to demonstrate value early, which helps maintain stakeholder support for the ongoing investment.

Get honest about your data foundations

If your underlying data infrastructure isn't ready for embedded analytics, no amount of front-end polish will save you. Before committing engineering resources to the analytics layer, take a hard look at the state of your data pipelines, schemas, and governance.

This isn't the most glamorous work, but it's the work that determines whether your analytics experience will be reliable, performant, and trustworthy. 

Skipping it – or underinvesting in it – is one of the most common reasons implementations stall or underdeliver.

Plan for the full lifecycle, not just the launch

Our data shows that the challenges of embedded analytics don't disappear after launch. Performance, security, feature requests, and tech stack evolution are all ongoing concerns. 

If your business case only accounts for the initial build, you're setting yourself up for a difficult conversation six months post-launch when the team needs continued resources.

Build your case – and your roadmap – around the full lifecycle. Treat embedded analytics as a product within your product, with its own maintenance, iteration, and evolution needs. That framing helps set the right expectations internally and avoids the common pitfall of declaring victory at launch, then slowly watching the capability degrade.

Address security and compliance early

With 42.9% of adopters citing data security as a key challenge, it's clear this isn't something you can figure out after launch. Multi-tenancy, role-based access, data residency requirements, and evolving privacy regulations all need to be part of the architecture from day one.

Adding security into an analytics experience that was built without it is significantly more expensive and disruptive than building it in from the start. Involve your security and compliance teams early in the planning process – they'll thank you later, and so will your users.

Consider what you build vs. what you buy

The technical complexity barrier is a big reason the build vs. buy question is so central to embedded analytics strategy. Our report found that 78.9% of product leaders say AI has changed their approach to this decision, with many shifting toward building internally for greater control – but others adopting hybrid models that combine internal development with external solutions.

There's no universal right answer here. But the data does suggest that teams who try to build everything from scratch tend to underestimate the ongoing cost, while teams who buy a complete solution may find it difficult to differentiate or integrate tightly with their product's unique workflows. 

A hybrid approach – buying the infrastructure layer while building the product-specific experience on top – is worth serious consideration.

The barriers are real, but they're navigable

Here's the encouraging part of the data: despite all of these challenges, 53.8% of products already have embedded analytics deployed or in beta, and another 27% are building or planning to. Teams are finding ways through – not by ignoring the barriers, but by being realistic about what's involved and planning accordingly.

The product teams that succeed with embedded analytics aren't the ones who find a way to avoid the complexity. They're the ones who acknowledge it, scope their efforts appropriately, invest in the right foundations, and treat the capability as an ongoing product commitment rather than a one-time feature launch.

That's not a particularly flashy takeaway. But if the data tells us anything, it's that embedded analytics rewards discipline more than ambition.

More embedded analytics insights...

This article is based on findings from the Embedded Analytics Opportunity 2026 report, produced in partnership with ThoughtSpot.

Download the full report for the complete dataset, expert commentary, and analysis across adoption, AI, and build vs. buy decisions. 👇