Problems with Automating Dashboard Reports

Automation is key to running an efficient marketing team or agency. However, with any automation there are limitations and issues that arise with it. As a marketing analytical company, we have automated performance reports, budget/pacing reports, sales reports and even project management reports. We have seen all sorts of issues that happen with automation, while 90% of the time the dashboards are amazing tools that save teams time and in most cases money, things can go wrong.

Reasons to automate dashboards:

Automating reporting is an efficient way to increasing performance, reporting results and managing budgets. Reports should be useful and accurate. A good first step to starting to build automation is to ask your marketing team or agency what they are currently using excel for and/or what manual calculations they are doing. Chances are that 99% of the analytical work can be automated where the dashboard is helping deliver insights to push the business forward. When building or designing your dashboard, the first step is to ask why? Why am I building it. What is the purpose of this metric or dimension, is it providing insight? You also want to make sure you are not overloading the dashboard with too many drill down and filters. A properly designed dashboard with the right stakeholders, KPIs and quality control section can help ensure that your dashboards are adding value to the team.

 StrategicDB, a marketing analytics firm has put together a list of problems that arise from automating reports and how to overcome these issues.

1.Broken Data Connector – When data is connected automatically via connector or direct API, it is great because it gets updated without any individual involvement. However, what happens if it breaks? And worst if it breaks without marketers realizing? Then you have incomplete data at best and at worst, you make the wrong budget decision for pacing.

The solution: add a quality control section to your dashboard where you can monitor the last day that data was added to the dashboard. You can also use max date and filter to show any data connectors that failed. Having your team, monitor the quality control section is going to ensure they are aware of any failures and can fix it before making any reporting errors. Depending on the tool you are using, you maybe able to use notifications to notify when data stream failed, this option is available on tools such as Datorama.

2. Permission Issues – Security may not be a big concern for some agencies, while others specifically in financial or healthcare settings, permissions and data protection takes a new meaning. Most dashboards have security implemented where only certain users with passwords and 2 factor authentication can access data. However, what happens when parts of the dashboard can be viewed by some team members and others cannot. Or if your business intelligence tool comes with limited number of users, how do you share it with your clients without encountering a larger fee. And finally, what if your analytical team does not have access to the raw data (or third party tool) that they are trying to implement. All these case scenarios require different approaches and would depend on each organization.

Possible solutions: depending on your agency you may have different work-arounds to the permission issues but here are some ways that these can be solved. Creating two dashboards based on user roles by duplicating a dashboard and removing sensitive parts depending on user permission. Creating different user permission groups and making them as “viewers” can help control who see what. For those companies who have limited access to data sources, you can do a screenshare where the people who have access input their credentials directly without analytical team getting access to third party tools. Alternatively, sending snippets of data to the dashboard based on the need to know basis, can help overcome some of the challenges. Finally, for user limitation, sharing the report as PDF while not ideal can be a work-around along with a weekly or monthly meeting to go over the dashboard in real time. These are all not ideal, but can be overcome. The importance is to know what limitations exist and ensuring that no data leaks happen while everyone that needs access has it.

3. Lack of Usage – At StrategicDB we are often told by IT teams or analytical departments at marketing agencies and corporations that all dashboards were built but no one is using them. This is a big problem because resources were used up to build those dashboards, while marketing teams are still pulling analytical report manually with pivots and PowerPoints. Most often then not, the real reason why dashboards are not being used is due to the fact that they are not useful to the department that asked for it.

To make dashboards more useful, it is important to ask requirements from Marketing. Displaying data that is available in marketing platforms will not make them use your tool. Adding value, understanding how they are currently using data and streamlining their current processes can help identify how the dashboard should be build. For example, displaying data per platform such as impressions, clicks, ctr, and media cost, cpc and conversion by campaign is not really adding value. Instead, showing how CTR or conversions have increased or decreased compared to previous month is your value add. If the team currently groups different campaigns together such as Top of the funnel, bottom of the funnel and mid funnel, then having a dimension that groups campaigns, provides them with the data that they are currently using pivots for. Finally, creating dashboard to copy their current PowerPoints and allowing a section for the team to update on commentary.

4. Lack of Trust In Dashboard: The second complaint we hear from marketing teams which have automated dashboards is that they have no trust in the data. This can be due to the integration errors of data, the fact that dashboard is missing critical information or that marketing agencies are adjusting their data from platforms. For example, some campaigns are run by different teams that are often excluded when reporting on monthly performance, while dashboard may not have those exclusions. Another example, is that there is extra cost that the team adds to their media cost when presenting to a client.

Solution to lack of trust is to first check all connections, secondary is to ensure that missing data is either documented or better yet, added to the dashboard. Once the data is all in and there are no missing data the second part is to understand what the teams are doing once they download platform data. And updating dashboard to exclude campaigns, add cost where needed or add custom calculations and other customizations. Getting the team involved in making sure quality control is done will help gain faith in the dashboard.

5. Lack of commentary: Sometimes automation is not telling the full story. For example, impressions maybe down this month, this could be normal due to seasonality. Another example in need for commentary is when additional insights is derived from a drill down analysis which does not make sense to automate on a regular basis. For example, the team maybe ran a segmentation analysis that enabled them to understand what products are sold best at what region to have a regional approach to the eCommerce site. While, sales by region is available displaying sales by region by product is not a useful way to display this information, which does not change often.

Solution is to allow the team to add a commentary section to the dashboard to provide additional context, insights and to identify areas that need it.

6. Too many filters, metrics, data: Sometimes analytical teams prefer to load every single point of data and be able to answer any question that may come up. This often overwhelms the user, where they are not longer able to see the big picture. For example, if there is a table that has a campaign, ad set, and creative and its performance all in one time, with a drop down by region, date range, product, demographic information, the dashboard is a copy of the actual platform data. This does not give any new insight to the user, who still has to drive the dashboard to be able to answer certain questions.

The solution to overwhelming dashboard is to stick to the top KPIs which are consistent. Showcase the data points that are needed for the team to move forward. If there are no action that can be taken to the team, do you really need that information? For example, it maybe great to know that cpc is higher for women then men, but what can the team do about it? If the target market is women, there is very limited action that they can to make the cpc match that of men. Therefore, sometimes less is more. Make sure that the dashboard has the main data points that are needed and keep it consistent, you can always create another dashboard for the drill down (in-depth) discovery dashboard which allows analysts to dig deeper and answer some other questions.

7. High Maintenance: high maintenance cost could include time that it takes to load the dashboard, regular updates that are needed from analytical teams and/or there is a high cost associated with your BI tool. If the team has to download data and then upload to BI tool on a regular basis or an update is needed from analytical team on a regular basis is dangerous because it opens up the dashboard to human error. It is also becomes inefficient after all why pay for a tool where the team has to do most of the work? And sometimes the cost of a tool is higher then having your team do manual reports in excel/PowerPoint based on their usage.

Solution: Identify ways to remove the human aspect of dashboards. Building custom APIs, finding third party connectors and automating/adding new process to your current dashboards can help streamline your dashboards in the way they should be operating. If it is not possible, it could be that you are using the wrong BI tool. For example, if your BI tool is missing Salesforce connector and that is the source for your Revenue numbers, it is time to find a tool that does connect to Salesforce such as Datorama. If the cost of the tool is too high for the amount of value it provides, time to find a different tool. Such as Google’s Looker (DataStudio) which is a cheaper alternative.

While automation is great, ensuring that your dashboards are correct, useful, and visible to the right group is key. Successful Dashboard starts with the right tool, design, implementation and quality control, it should not require a lot if any on-going maintenance and should be useful to your marketing team.

If you are struggling with your current dashboards, or want to automate marketing dashboards, StrategicDB can help, they are a full service marketing analytics company that specializes in helping automate dashboards for marketing agencies.