8 Avoidable Mistakes That Are Killing Your BI Projects

The future present is data-centric, and for businesses to compete in this ever-evolving – not so byte-sized – digital economy, they require insight into their data and analytics to assert themselves as players for years to come. Enter, Business Intelligence (BI). Business Intelligence – data analytics, reporting, data visualizations, etc. – is routinely a hot discussion topic; one that continues to gain steam and is showing no signs of slowing down. Once seen as niche, and scarcely exploited, BI has since proliferated, becoming a necessity for businesses looking to outperform their competition.

However, where Business Intelligence has the ability to be the vehicle that transforms a company from one that merely survives to one that thrives, the road to data-driven glory is paved with the remains of failed ventures and ill-conceived strategies. BI can help businesses immensely – from identifying mechanisms for increasing profits to predicting the success of future projects, to analyzing customer behavior – but only if set up properly… the first time.

Business Intelligence projects are complex and often not straightforward. They’re ever-changing; a living organism that should be treated as such. The majority of BI initiatives are vast undertakings involving several parties; sometimes they result in success stories, sometimes they fail. There’s no magic wand or quick-guide to ensuring your particular endeavor is a roaring success.

There are, however, several common mistakes we see being made in unsuccessful BI projects time and time again. To ensure your Business Intelligence project is a success, I’ve outlined 8 common mistakes businesses make (and how you can avoid them). I’m not pulling any punches; you’re about to dive into a detailed article on BI pitfalls. But don't worry, we've made it easy to navigate and jump through the various sections. AnchorTop


8 Avoidable Mistakes That Are Killing Your BI Project

  1. Tracking & Implementing the Wrong KPIs
  2. Inability to Tame Unruly Databases
  3. Underestimating the Importance of Product Flexibility
  4. Lack of Communication Between the Business & Technical Sides
  5. Confusing Self-Service with Self-Sufficiency
  6. Not Enhancing your Analytics with Data Visualizations
  7. Inability to Communicate the ROI of BI / Get Buy-in
  8. Inability to Instill a Data-Driven Culture


If you can’t measure it, you can’t manage it” – or so the old adage goes. This is why Key Performance Indicators (KPIs) are so important. It’s essential that businesses are able to communicate their objectives across the organization; KPIs ensure these objectives are top of mind and are critical to helping properly track, understand, analyze, and utilize data.

These indicators should – by all means – be the most important measures/units/values businesses use to track their progress against said strategic objectives. Well-designed KPIs are akin to lighthouses; they are pivotal tools for navigation and provide guidance on where businesses should – or should not – be heading. If properly established, KPIs will improve performance and provide clear guidelines on how to measure progress.

However, KPIs – and Business Intelligence projects by extension – are prone to failure, as they’re often poorly-defined. Most businesses mistake KPIs for Key Results Indicators (KRIs). While both are paramount to day-to-day monitoring, analysis, and performance measurement, KRIs merely track and define success. KPIs, on the other hand, break it down and measure it. Take, for example, a marketing dashboard containing the following indicators:

> Bounce Rates
> Page Views
> Lead Conversion Rates
> Average Time on Page
> Etc.

These are KRIs. With these, you’d certainly be able to deduce the status of your – for example – content strategy. You’ll know whether these indicators are performing to expectation, and you can adjust your processes accordingly. However, you won’t have a sense of how impactful your changes are until some time has passed and your dashboard is updated with newer results.

To rely solely on KRIs is to be reactive; they’re useful for understanding what’s been done. Take, for example, a second dashboard. The difference between this and the first, is the indicators being tracked now, are ones that will lead to the results you want:

> Meetings Booked
> Proof-of-Concepts Won
> Demos Completed
> Etc.

These KPIs are tracking what needs to be done in order to meet those pre-defined strategic objectives. With KPIs, you’re tracking critical success factors, in an established process, in real-time, to ensure you’re heading in the right direction. Don’t make the mistake of mindlessly chasing numbers.


A poor database can be the bane of a Business Intelligence project. This is a database that’s messy, outdated, formatted incorrectly, and on occasion, suggests the improbable – no, that customer has not been active since the 1700s. The odds of your database being perfectly suited to building dashboards and reports are slim to non-existent; it’s more probable they resemble trash cans. I daresay finding a pristine database is less likely than finding a unicorn; sometimes I question their existence.

Let’s get one thing straight. By no means are databases intentionally unruly. It’s something that happens organically: short-term decisions often lead to questionable design choices, data gets hoarded, naming conventions go the way of the dodo bird, business processes don’t change over time, etc. Plus, the inherent messiness of data plays a pivotal role; data, more often than not, is incomplete, nonstandard, inconsistent, inaccurate, and dated.

Let’s take, for example, a Customer Relationship Management (CRM) platform. The success of a CRM database hinges on the interdependency of the parties using it; Sales, Marketing, Customer Success, etc. Despite each team’s best efforts to be communicative and united, it’s safe to assume each time a change to the database is made – such as the addition of a new field – not everyone is made aware of it. This is to be expected, as databases are constantly rewiring to adapt to changing business needs, however, these minor tweaks and modifications to databases compound over time, leading to some perturbing consequences.

So, instead of focusing your efforts on addressing data and database quality (both of which – don’t get me wrong – are vitally important), consider instead, selecting a Business Intelligence platform that intrinsically possesses the functionality to remedy an unruly database – right from the onset - such as Dundas BI.


It’s my assumption that you are a stellar evaluator and are thorough in every aspect of your Business Intelligence evaluation. I’d even wager you put Ford’s 172-point inspection to shame! The thing is, regardless of how immaculate your assessment might be, your current BI and data visualization requirements today will not be your requirements in the future. This means that even if a BI tool checks all your boxes right now, that likely won’t be the case at some point down the road.

So, what can you do to stay abreast of your fluid, ever-changing requirements? How can you ensure your BI tool is adaptive and can keep up with whatever’s thrown at it – or will be thrown at it? How can you make worrying about the future, a thing of the past?

If you’re evaluating BI tools there’s one criterion the tool you select must meet, that’s head and shoulders above the rest: flexibility. You need to make sure that the tool you ultimately select will be able to grow and expand as your requirements grow and expand. I urge you to be particularly diligent in making sure each tool you evaluate is flexible in all areas of the application. In fact, I’m so adamant about this commonly avoided mistake, I’m going to share with you a list of questions you should ask each BI vendor when evaluating the flexibility of their tools:

Data Connectivity > Which data sources do you support? > What about data sources that aren’t supported out-of-the-box? > Do you support multi-dimensional/OLAP data?

Data Preparation
> What data storage techniques do you support?
> Do you support user-based security?
> Will your tool clean, transform or augment my data?
> Do you support data governance?
> Do you support data correction?
> Is writeback and data input possible?

Data Analytics
> Can I transform data with Python, R or scripts?
> Do you support on-the-fly what-if analysis?
> Do you have a plug-in model for adding new formulas?
> Do you support ragged hierarchies?

Data Visualization
> What advanced data visualizations are available?
> Can I add my own data visualizations?
> Do you support mobile/responsive design?
> Can period-over-period be easily added?
> Are report bursting and notifications supported?
> How does embedding work?

> How are development methodologies supported?
> Can the administration be done via API?
> How can I configure the application?

… This is just a brief extract from an exhaustive list of questions you should ask. If you’d like the full list of questions with a more detailed explanation, please refer to this video >>>

These questions are meant to challenge technical products. If the Business Intelligence vendors you’re evaluating can answer these questions easily, with proof, then you’re on the right track. The most important question you need to ask in regard to product flexibility is ‘how’. If there’s no proof in their pudding, buyer beware.


In any Business Intelligence evaluation, there are two sides: the business side and the technical side. In many instances, the business is the what and the technical is the how. Unfortunately, while both sides have the business’ best interests at heart, each has competing perspectives and beliefs regarding the function of BI and how it can and should best be utilized. Complicating matters further is that the business side tends to be overly involved in the wrong stages.

To ensure your business intelligence project succeeds, it’s imperative the what and the how come together. It’s well known that Business Intelligence is a business driver – empowering businesses with the ability and opportunity to make more informed, data-driven decisions – however, the best BI deployments are achieved when the business and technical sides work cross-functionally and collaboratively, to achieve a single goal.

OK, great, so we’ve identified that both sides need to work jointly for a BI project to succeed, but what does that mean exactly? Well, most businesses fail to understand or communicate the behavioral outcomes that a BI output (i.e., dashboards or reports, etc.) needs to provide. And it’s a lack of communication between the business and technical sides that create and subsequently broadens this rift. To remedy this – or better yet, to prevent this entirely – the business side needs to clearly and concisely state – from the onset – their desired outcomes along with why they’re necessary. Too often, technical resources are left to develop in a vacuum and don’t have a concrete enough understanding of why they’ve been tasked with certain requests.

This topic is nuanced and complex, but what I recommend, is starting from the end and working your way backward. By beginning with the business side and their objectives - who will be looking at the data, why are they looking at the data, what outcomes are they expecting, etc. – you’ll be able to better understand and define which visualizations and output make the most sense. From there, the technical side can traverse back to the data and prepare everything accordingly.


Now more than ever, businesses are embracing the idea of self-service analytics. In fact, this trend is so popular, that Gartner has predicted that by the end of 2019, “the analytics output of business users with self-service capabilities will surpass that of professional data scientists” and that by the end of 2020, 80% of all enterprise reporting will be comprised of self-service Business Intelligence tools. Great, so self-service analytics is the future, right? Wrong. Based on our experience and the projects we've worked on, only 10% of business users actually need self-service capabilities.

The idea of self-service analytics is alluring. It promises to free analysts so they can focus on more strategic and valuable work other than reporting, and it promises to empower business users with greater access to their data, whenever they need it. Businesses are tired of the rigidity and lack of intuitiveness that legacy, traditional reporting tools possess, and are understandably seduced by the power and freedom that self-service pledges. But is their focus misplaced? Does self-service over-promise and under-deliver? Herein lies the problem; the mistake that businesses often make is confusing self-service with self-sufficiency.

Everybody thinks they want self-service, but that’s not necessarily the case. Does everyone actually want to blend data with new sources on-the-fly? Do most end users even know what data blending is? And even if users want to perform more advanced data preparation, do they have the time? The answer to all of these is a resounding ‘no’. What businesses truly want, is self-service on a silver platter – or as we at Dundas like to call it, silver-service. This variant on self-service analytics is achieved when the data business users need to perform their jobs is gifted to them on a silver platter – hence the name – in a manner that’s intuitive and easy to consume. The method ensures business users aren’t required to learn how the software works yet are still able to explore data in a tailored environment.

Self-service is enticing. That said, it’s not a means to an end; it will not take your business users and transform them into table-splicing, data-cleansing, model-building analysts. During your evaluation, don’t make the mistake of falling for the empty promises of self-service analytics and instead, look for a BI tool that is designed with all user types in mind. The tool must empower analysts and BI professionals to dig even deeper into their data and perform in-memory or on-the-fly data analysis using highly interactive visualizations and an equally powerful analytical engine. In turn, this group of users needs a tool that allows them to turn their findings into beautiful reports that can be shared with the majority of users in a way that works for them. Business users must then be provided with a fully governed experience that allows them to personalize these reports and create relevant and actionable content that is easy to consume, and more importantly, easy to use.


Business Intelligence is a catch-all term, in that it comprises many different elements from the strategies and technologies enterprises and software vendors use for analyzing data. One of those elements – which I’d argue is the most important – is data visualization. The fact of the matter is, companies are drowning in pools of data; there’s simply no shortage of valuable information at their disposal. But all those 0s and 1s are worthless, without the proper tools in place to understand them. Here’s where data visualizations come into play.

The presentation layer of your Business Intelligence and analytics stack is your window into your business’ performance. By using data visualizations, you’ll be able to drive instant decision-making – by turning massive, raw quantities of data into visual insights – which will ultimately lead to better results. You’ll also be capable of consuming data more effectively, leading to discoveries you’d previously never have expected to unearth.

We’re naturally hard-wired to process, interpret, and respond to visuals faster than we are rows of upon rows of data. After all, we do come stocked, out-of-the-box, with the most advanced pattern recognition technology on the planet – our brains. This is why it’s of the utmost importance that data visualizations should take precedence when analyzing data, to ensure everyone has a complete picture – pun intended – of their data.

The best Business Intelligence vendors are the ones that understand the importance of stunning data visualizations. And the right business intelligence tools are the ones that drive greater data literacy through visualizations. Anyone can slap a data visualization on a dashboard or report and call it a day, but the cream of the crop knows that effective data visualizations are highly customizable – right down to the nitty-gritty details – flexible, beautiful to look at, and interactive.


“Information is the oil of the 21st century, and analytics is the combustion engine”Peter Sondergaard, SVP Gartner, 2011. While this metaphor isn’t perfect – as there are categorical problems with treating data as a natural resource – it clearly encapsulates the value of data and the importance of making sense of the plethora of it at our disposal. And what better way to make sense of data, than with Business Intelligence? BI is – as any BI vendor will gladly state – an invaluable asset. It’ll improve efficiency, assist with identifying new growth opportunities, and ensure you’re equipped with the insights you need to make better, more timely, accurate business decisions. It’s easy to list BI’s innumerable benefits. What’s less easy to do, is pinpoint it’s ROI and get internal buy-in; and here’s where a lot of BI projects derail.

The total spend on Business Intelligence tools is already astronomical, and its upwards-trajectory is simply mind-boggling. In 2016, the market was valued at just over 17 billion, and by 2025, is projected to reach nearly 150 billion. That’s a lot of money to put towards something with a reputation for being notoriously difficult to value. Cleary, Business Intelligence is a game-changer, but unless you’re able to quantify its value and get everyone on board, you may as well quit while you’re ahead. Too many BI projects fail before they even begin because the project leader is unable to communicate its ROI.

So, how can you determine the return on investment of Business Intelligence? Unfortunately, there’s no standard answer to this question as the benefits to one business will not be the benefits to another. That being said, we believe every successful BI project begins with a definition of what success looks like. It cannot be stressed enough, but it’s not enough to simply visualize data; there must be processes in place that ultimately strive to change behavior. Once you understand what exactly is it that you’re trying to achieve with BI, you’ll find the actual valuation of the investment that much easier to obtain.

Here are more detailed, specific examples of what BI brings to the table – at both the department and industry levels. Use these as a foundation for understanding and calculating its expected value.


At the end of the day, your Business Intelligence project will only be a success if your users actually use the tool! And even then, simply using the tool is often not enough. You need to instill a data-driven culture; one where data reigns supreme, putting the entire business in a position to make better, more impactful business decisions. Not doing so puts your entire project in jeopardy, leaving it precariously perched, teetering on the edge of failure.

There are numerous ways of creating a more data-centric culture. It’s not something that will happen overnight – it’s an investment of time, effort and money – but with a little creativity and elbow grease, it’s just a matter of time before data is used pervasively through your business. Here are a few suggestions on how to install a data-driven culture:

Data Flows from the Top, Down
If a leadership team is dedicated and involved in creating a data-driven culture, you can bet your bottom dollar their enthusiasm will have a cascading effect. You’ve heard the phrase “leading by example” and this is no exception. Your leadership team must exemplify the traits you wish to imprint and must immerse and commit themselves to be data-driven at every stage of the process. If your leaders aren’t loyal to the cause, well, then data-be-damned.

Make Data Accessible for Everyone
Business Intelligence is complex enough; don’t over-complicate it for your users by making it needlessly difficult to access and use. Accessibility means bringing easy to understand data into the tools and applications your users are already familiar with. One of the best ways to do so is by selecting a Business Intelligence tool designed to integrate with and embed into existing systems – doing so puts data on everyone’s radar.

It also doesn’t hurt to double-check how proficient and versed your BI tool is in the realms of communication and collaboration. Features such as data-driven notifications and scheduled reports go a long way in creating a more data-driven culture.

Educate and Support New Users
Like any new initiative, success hinges on adoption. Take time to train and educate your users on the value of BI and how to use the tool. Give your users everything they need to be successful and don’t be afraid to leverage the support BI vendors offer. Don’t make the mistake of assuming your users will immediately adopt your solution. Make Business Intelligence habitual and your project will be immensely successful.


Make No Mistake About It

Mistakes are bound to happen – it’s the nature of the business. At the end of the day, the best you can hope for is to limit the aftermath of each one made. It’s impossible to account for every scenario. Preparation is key when planning a Business Intelligence project, and a well thought out strategy will ensure the end result is well worth the investment.