• Navid Nassiri

Five data integration pain points (and how to avoid them)

Many digital first companies are looking for more efficient ways to manage their data. Well managed data allows teams to extract actionable insights about the metrics that matter to them. They get to know their customers better, and in turn, increase revenue. Still, some companies are struggling to keep up with their data challenges.

So, what’s the hold up? The biggest issue is the soaring volume of data. But there’s also the fact that many of the legacy systems on the market weren’t built to meet today’s three-fold challenge: 1. aggregating different types of data, 2. doing so at an unprecedented scale, and 3. being able to troubleshoot without burdening the tech team.

Here are some of the most common data integration challenges we see companies facing every day, and how they can avoid them in the long term.

Pain point 1: My current data management solution is getting the job done, but the performance is suffering as we continue to scale.

How to avoid it: There are several existing systems built to aggregate media revenue data. The trouble is that daily volumes have exploded from gigabytes to terabytes – even petabytes. Some platform providers are beginning to innovate and advance their technology. But there is still a great deal of work to be done in the industry to keep up with users’ needs.

It’s no longer worthwhile to battle with over-burdened platforms that will leave you behind the competition as data volumes continue to grow. Now is the time to look for data engineering automation platforms that can already handle data at scale. They also must be agile enough to grow with your business.

Pain point 2: The data I’m aggregating is still stuck in silos. To extract meaningful insights, I need a unified dataset that scales across the organization.

How to avoid it: Some data aggregation tools are great at processing one type of data – for instance, programmatic revenue data. However, the recent digital surge, accelerated by Covid, means media teams are expected to sell more inventory, including direct deals. Therefore, having a blended view of your inventory is critical.

If you’re struggling to combine programmatic revenue data with direct deal data, you’re not alone. In fact, even after an unsustainable amount of manual reporting, the result, for many of the customers we talk to, is a poorly presented bar chart which is dated the minute it is produced. Instead, look towards solutions that can blend a number of different datasets – automatically – and integrate with sophisticated data visualization tools.

Pain point 3: My current platform underperformed on the implementation. Instead of the month promised, it took several months and still is not 100% ready to use.

How to avoid it: First, platform providers should manage expectations so that business teams know how long the implementation phase is likely to last. This will not only prevent disappointment, but it will allow them to plan around the set-up stage and allocate enough internal engineering resource to handle data aggregation and reporting needs while a new platform is implemented.

Second, teams should check in advance to make sure any platform they implement is agile enough to stand up a full reporting system – as well as create bespoke “data recipes” (or normalization templates) – in the time frame required. And not incidentally, this should be a matter of weeks rather than months or quarters.

Pain point 4: Using our current platform requires continual involvement from our engineering team. Access to data is restricted to technical teams as there is a code-barrier to entry that prevents our business teams from gaining timely, actionable insight.

How to avoid it: One of the most common complaints we hear is that business teams are prevented from accessing the data they need because combining new and/or changing data sources, or extracting new insights, requires coding skills. Some platforms are unable to support additional connectors automatically. So, in-house engineering teams are pulled in to write the code, which in turn diverts them away from other business-critical initiatives.

These teams need an ongoing data partner rather than simply a vendor who hands over the credentials and leaves them to it. A data partner should be on hand to help business teams create automated self-serve templates whenever they need a new slice of the data, and to build new API connectors for any new or changing data sources. Before implementing a platform, be sure to check whether these processes are automated.

Pain point 5: Our current platform takes one to three weeks to escalate or acknowledge even simple tickets.

How to avoid it: Accessing or managing company data isn’t a nice-to-have – it is a business imperative and can impact critical decision-making. In our experience, some platforms have become so large that they are unable to prioritize tickets effectively. Instead, business teams should insist they have their own dedicated Customer Success team. Data integration issues can be triaged almost immediately, which ensures that pressing issues can be solved (or at least scheduled to be solved) in a timely manner.


Providers should state how quickly they will respond to each type of query in their SLA. Look for one that will respond within the appropriate time frame depending on the nature of the ticket.


In a world where data volumes will only continue to surge, now is the time for data engineering automation platforms to step up and make sure they can handle any type of data, at any scale, accurately and reliably.


Digital first companies need to take control of their data ecosystem. To achieve this, they need to make sure they ask the right questions when making the ‘buy versus build’ decision and placing their trust in a long-term data partner.The resulting new data-driven capabilities and insights will drive real engagement with their customers, and make a real impact on the bottom line.


Read more data-related blogs HERE