Here’s the communications dashboard development process broken down into six steps:
(Please note: If you want some background knowledge, I recommend checking out these case studies about dashboard implementations: “How to Cure the Sick Dashboard of a Big Healthcare Company: Step-by-Step Intensive Care and Rehabilitation” and “Build a Measurement Dashboard Now & Survive a Crisis Later.”)
Step #1: Audit, agree on objectives, and decide on metrics (Time required: 1 month)
The first step in creating any dashboard is a thorough audit of research and evaluation processes within your organization. This can frequently yield surprising results. In organizations where there is little or no research at all, you get to start from scratch and do it all correctly, but that is seldom the case. For example:
- One senior VP who was starting down the dashboard road asked her staff to bring her copies of all the research and evaluation reports her department was receiving. There were 36—and of those 36, only about 5 were delivering information that anyone trusted or was using.
- In a destination marketing organization, the CMO, a very research-minded manager, had terrific ROI statistics on everything in his organization—except PR and social media.
Once you understand the state of measurement within your organization, the next step is to agree on objectives. There are two sets of objectives:
- The first set of objectives are those for the dashboard itself: What is it supposed to achieve? What value is expected from the implementation?
- The second set of objectives are those that are fundamental to your business or organization: What are the business objectives that your dashboard will be measuring against? These objectives will drive the metrics that will appear on your dashboard.
Once you’ve agreed on those two sets of objectives, you can develop your implementation plan. Allow at least one month to develop your plan. As you do, make sure you set expectations correctly.
Allow plenty of time for consensus-building and deciding on metrics. Dashboards are complex. If yours is going to be used as widely as it should be within your organization, then you’ll need a lot of people to sign off on your metrics and the overall plan. Here are two examples:
- For one travel destination, with a communications staff of less than 10 people, the process was relatively simple. We were able to get the executive director in the same room with the staff and a few of the agencies. We had a productive discussion and agreed that getting traffic to the “Visitor Guide” page on the website was an acceptable proxy for intent to visit. They were already using quarterly survey research to measure actual intent, awareness, and preference. So we inserted a question in the survey to identify people who had heard news via earned media. We were then able to track the extent to which people who had seen destination-related news in key markets were influenced to perceive the city as a fun place to go with friends — a key driver of actual visits.
- On the other hand, for a major high-end retailer, the meeting to define the metrics was contentious and took a full day. In the end, the only outcome metrics they could agree on were the percentage of items that contained a desirable visual, key message, and call to action. The resulting dashboard was very simple.
Step #2: Decide on your dashboard platform (Time required: 2 weeks)
There are any number of different dashboard platforms out there. Every major tech software company has one: Oracle, SAP, SAS, IBM, etc. So if you work for a large organization, there’s a good chance that there’s already one in use. Check with IT and business intelligence to see what they’re using and how you can get access.
Some of the more sophisticated measurement platforms, like CyberAlert, Prime Research, and Talkwalker, also provide dashboards that can integrate other data into their existing structure. If you’re thinking of adding or upgrading your media or social listening system, and there’s a dashboard in your future, you might want to start by talking to one of these.
If you have to start from scratch, my advice is to first build your dashboard with PowerPoint and Excel. Here’s an example of a PowerPoint dashboard:
While dashboard systems like Tableau are wonderful and powerful, they need specialized knowledge. Unless you have an expert in-house, start simple and work up to fuller automation once you’ve got all the kinks worked out. If you do decide to go with a new platform, test several to make sure they’ll do what you need them to do.
Step #3: Source your dashboard data (Time required: 1-4 months)
You’ve already completed Step 1, above, so you should know where your data is coming from—maybe. In many organizations everyone knows data is available, but no one knows who exactly has it. Therefore, you may have to do some digging. Send out emails to the keepers of the data, and follow up with an in-person visit. Buy lunch if necessary.
One thing to watch out for is the “information is power” syndrome. Some people guard their data the way dragons guard gold. It may require the power of the CMO to get people to hand over data.
In government institutions or highly-regulated national organizations, it is very likely you will hit an institutional brick wall. Or there may be an assumption that data exists when in fact it has never been collected. In several organizations we encountered vendor contracts that hadn’t been signed (hence the data wouldn’t be available for months), or data sources that were once available had been eliminated for budget reasons. That sent us back to the drawing board to change, replace, or eliminate specific metrics.
If there are gaps between what you need and what you have available, then you may need to put out a request for proposal (RFP) to find a research supplier to provide the data. This may require competitive bidding. Keep in mind that if you hire a supplier tomorrow, it will still take 2-4 months of research, testing, validating, and cleaning before you have usable data (see “6 Companies Crawl to Barcelona Compliance: Part 2, Getting Good Data”).
Step #4: Validation and testing of data (Time required: 1 month)
Once you have the data, make sure it is all in Excel or another format that is compatible with whatever dashboard platform you’ve selected. Then make sure it is valid and complete. If you are using sentiment and message data, then it needs to be checked for accuracy, no matter whether it is human-coded or machine-coded.
The next test is completeness of data. If you’re starting with a six month benchmark, do you have data for all 6 months? Is the data consistent in its time frames? (Web data is typically delivered by day, media data is typically done by week or month.) You will need to spend time reconciling dates to make the data usable.
If you are using an aggregate index, this is the time to test it to make sure it makes sense. Enter real data and do the calculations: Are they reasonable? Useful? Credible? If not, then your index may need tweaking.
Step #5: The Real Thing (Time required: 1 month)
Once your data is clean, in the right formats, and flowing into your dashboard you’ll be able to create the actual visual representations that will be popping up on your dashboard every day, week, or month. Warning: Getting this part right is harder than you think.
You’ve probably drafted your dashboard in PowerPoint, but that has little to do with how a dashboard actually looks. PowerPoint is a static chart that most people print out and read on a plane. A dashboard is intended to be a dynamic visual presentation that changes in response to new data or reorganizations of data.
So the next step is to make sure that all the necessary elements are included, so that everyone can make sense of the charts being presented. For instance, use headlines that describe the chart accurately, legends that are legible and understandable, and charts that are readable.
Unless you want to deal with endless tweaks, try to get everyone who will be using your dashboard in the same room to review things at the same time. This way all the changes can be recorded and implemented at once.
Step #6: Putting your dashboard to work
As wonderful as a dashboard is in keeping data up-to-date and providing constant access to the information you need to make decisions, it is still just a means to an end. Measurement and evaluation should always lead to insight. You are the only one with knowledge of the data and its context, so you and your communications team need to provide that insight.
To make the best use of your new dashboard, convene a group of people you trust that understand what activities went into the results. Lead a discussion of what caused any changes, and try to understand why numbers went up or down. Rank order programs from “best performing” to “worst performing,” and discuss why some worked better than others. Then you can write the headlines, conclusions, and recommendations. Brainstorm on better solutions.
Here’s how some of your peers responded to their first set of results:
- For a travel destination: During a summer of overwhelmingly bad news, they used their data almost weekly to make decisions about how to allocate resources to mitigate the damage most efficiently. As a result, the PR budget was renewed while all other budgets were slashed.
- At a healthcare organization: There were several senior management changes during the roll out, and a period of unusually heavy media attention, mostly unwanted. As leadership shifted, different priorities emerged and metrics were adjusted in the process. However, the consistency of the data and standard reporting styles proved to be an enormous asset. Not just in the extent of the coverage, but also in what themes and messages helped mitigate the undesirable messages.
- A nonprofit organization: Their data was used to make resource allocation decisions. As a result they have become better stewards of their donor’s contributions.
Bottom line: A communications dashboard is a very useful tool to get the data into a format from which you can make better decisions. But don’t forget that the insight still has to come from you, yes y-o-u.
∞
