Much of Marketing is totally pragmatic. We actively look for a different way to view an issue, so that we can ask fresh questions, get different answers and create innovative solutions. Key Performance Indicators are classic example.
Here are nine thoughts on KPIs based on a variety of experiences across industries, processes and systems.
1 Part of the Big Picture
The purpose of KPI reporting is to support decision-making on critical issues.
The content is deliberately selective. It doesn’t aim to cover everything. Instead, it highlights key issues within a larger area. But what it does cover, it reports faithfully and without distorting the scale of the issue.
Reports highlight and summarise the essential issues accurately and concisely so that the Audience can understand the current situation. Their next step will be to ask questions and discover whether discussions are needed, whether decisions must be made.
2 Reporting as communication
When you stop and think about it, a KPI is a form of communication between a Reporter and their Audience. Ideally the reports will communicate meaning to the Audience without the presence of the Reporter.
For that communication to work successfully, the Reporter and Audience have to share a common understanding of:
- Context – what the report is referring to
- Content – what it means: why these units / this layout
- Implication – which decisions can be taken using this information.
There’s simply no way that happens by accident.
In my experience, it’s a lot more productive to create, review, and refine those sketches on a white board before diving into the details.
Reporter and Audience must sit down together to define the essential business issues; identify what business decisions are linked with those issues; and sketch out how the relevant information should be communicated.
3 Form follows Function
The idea that report content must be self-contained and the display intuitive is important. But in the initial design stages, I think the graphical issues around Design of Information and Visual Communication are a distraction. I especially like Steve Jobs’ quote about the iPod, because I think it’s also at the heart of the whole KPI thing:
“Design is not just what it looks like and feels like. Design is how it works.”Steve Jobs on the iPod
If we take functionality as our starting point, a useful first question to discuss with the Audience is: “what function should the KPIs perform?”
Here are a few thoughts on different types of report.
Status A status report shows how something was at a specific point in time (“high tide was 1,2 m above normal sea level”), or how it is now (“body temperature is currently 36.8 Celsius”). To be meaningful, the display should also indicate whether the value is acceptable or not.
Performance Revolutions per Minute. Kilometres per hour. Website visitors per day. Purchases per week. Counts, volumes, values. These are often a variation on the status report.
Trend How an effect (by convention the y-axis) changes with variation in a cause (the x-axis). A classic example is revenues over time (months, quarters, years).
A trend report is often a status or performance report that includes recent history (“we got here via this path”). The danger of trend reports is the temptation for eye and brain to think it sees a pattern and use them as a predictive tool – which they most definitely are not.
Alarm Bell The function is to trigger action, so there has to be prior discussion and agreement between the Reporter and their Audience. First about the sensitivity of the trigger setting and second about the follow-up procedure.
After all, there’s a serious difference between: “ten minutes after the alarm went off, the team had put the fire out” and “the building is burning because no-one knows what to do when the alarm goes off”.
Crystal Ball Reports that suggest how things may evolve in the future (i.e. time-orientation). For example: “there’s enough reliable historical data available to indicate, via sound statistical calculation, how a situation might evolve in the near future within defined confidence limits”. And yes, the data analysis does have to be that pedantic to be valid.
Radar Screen What’s coming into view (i.e. spatial / conceptual orientation). Vitally useful in business and yet tricky to do because there are so many eventualities to cope with.
You might for instance, devote resources to following the science of new materials to identify opportunities for creating new products. Or you might put resources into tracking the evolution of customer needs. But it’s only by doing both that you can create a new product which satisfies customer needs.
Progress We might want to show how far we’ve come (“Year to Date”); or we might want to show how much remains to do (“quantity remaining”). To run a quantitative analysis, we need a data source of units that can easily be counted, such as time, resources, money or whatever. If quantative sources are missing, or to do a qualitative analysis, you might show milestones instead (“Stage 3 of the project is now complete; only Stage 4 is remaining”).
Flashing Lamp An exception report. “Hey! We spotted something. It looks like this and we evaluate this way. Let’s discuss whether we need to take action … etc.”
4 Applying these ideas to Marketing
It helps enormously if Marketers begin by clarifying who their Audiences are and what information they need.
Cars provide an easy-to-understand example
- A car dashboard enables the Driver to operate the vehicle safely: the fuel tank is full, no warning lights are flashing, the seat belts are in use; etc.
- When the vehicle goes in for maintenance, the Mechanic accesses an entirely different set of (historic) data, from a different system, for in-depth analysis – and gives the Driver recommendations (“it’s time to change the brake pads”).
- The car navigation system enables the Driver to tell the Passenger how far it is to the destination, the expected time of arrival and whether to expect any disruptions en route.
In much the same way, Marketing departments need different types of report, according to the specific needs of the Audience. You get the idea.
5 Reporting by Marketing
In my experience, Marketing creates:
- reports about itself, for its own use;
- reports that communicate about Marketing to non-Marketers.
They’re very different in content and style.
Operation reports for Marketing Imagine for a moment Marketing as a ‘Black Box’. There are resources inputs (staff, budget, data) which are combined into processes (online, offline, etc.) to create outputs (events, leads, prospects).
The operation reports provide an overview of the health (effectiveness, efficiency) of that ‘Black Box’. They answer questions like. “are the operational systems and processes consuming the expected levels of resource and delivering the desired level of output?”.
Maintenance reports for Marketing These might be the reports from Team Leaders / Head of a function or system to the Marketing Director.
The top level is the summary of recommendations for action; below that is the analysis that explains the recommendation; below that, the details (once the method has been jointly agreed, the Marketing Director probably does not need to see these).
Navigation reports for Non-Marketers In parallel, there’s a set of reports that Marketing Directors want to share with Peers, other department heads and Stakeholders.
The purpose of these reports is to make sure that both left and right hand are working together to achieve business objectives. The format and content will depend on the audience: Sales, Finance, IT, HR, CEO.
6 A brief word on Data
It’s not always a good idea to begin with the data you’ve got, or what standard reports a system can spit out easily. Why? Because it limits your thinking.
Standard reports reflect an existing process. Using those reports the best you can do is optimise that process – which may actually be inefficient or unfit for purpose. A 1984 Rolls-Royce Corniche may be in tip-top condition – but the fuel consumption is unlikely to be better than 28 litres per 100 km (9 mpg UK).
Real improvement comes when you re-think how to achieve the objective and re-design the process. In practice, that often means collecting new data.
Defining what data you need to control processes (but don’t have yet) is itself a strategic objective. Rather than ignore it, you could put a hand-drawn sketch in the Report book labelled “feedback wanted” or “coming soon”, with a status update on the milestones toward delivery (data need defined / data capture started / report available).
7 Roll-up, Drill-down
Once you’ve outlined a report, you will probably need to think about data granularity. At what level of detail do you need to collect the data so that you can roll-up to totals, or drill-down to details?
The choices for data dimensions are almost limitless here:
- Time: seconds / minutes / hours / days / weeks / months / quarters / years
- Geography: ZIPcode / City / State / Country / Region / Continent
And in addition to the generic ones, there are all your company-specific dimensions:
- Customer segments
- Product / Service categories
- Internal organisation: sales regions, divisions, departments, etc.
As a side note: combining data from different systems can be very messy and resource intensive – but you probably knew that already.
8 Which subjects should KPIs cover?
Let’s go back to the opening comments. I suggested that: “KPI reports do not aim to cover everything. Instead, they highlight key issues within a large area”.
A useful question may be: “what criteria can we use to identify the key issues?”
I think that one set of considerations to use here is the ‘Black Box’ described above.
- which activities / processes get the most Inputs (staff / budget / IT / etc.)
- which activities / processes generate the most Outputs (volume / value)
In addition, I think a macro-scale version of the Navigation report may also be useful.
“How well is the company achieving its long-term objectives?”
To be able to communicate on that subject we need to know what the company objectives are and how the company defines the short- medium- and long-term. Direction, speed of progress and building new capabilities / infrastructures are thoughts that spring to mind here.
In this context, Marketing needs to communicate about things like:
- staff: development & retention, successor planning, building in-house skills
- systems: data collection, quality, integration, privacy by design
- processes: online and offline, how to join them seamlessly, digitalisation & re-design
- budgets: for ongoing operations / for investments in new capabilities, infrastructures
9 Looking to the future
One thing that can help Marketing increase effectiveness is to make a mental transition “from Project to Process”.
Many teams within Marketing tend to think of the year as a sequence of projects. For example: Online does a series of Lead Generation Campaigns; Events does a series of Trade Shows; Corporate Communication does a series of Articles; Product Management does a series of Brochures. Start a project, finish it; start the next one … Repeat.
Another way of looking at this is to say that each team manages a process that produces Campaigns / Events / Articles / Brochures / whatever. Each process combines a set of resources, skills, data, systems and budget in a specific manner to achieve a result.
So what would happen – bear with me, I’m thinking out loud, here – if we turn our Marketing professionals loose on those processes and challenge them to find new ways of combining resources?
Perhaps we could create a Trend chart showing “Total Resource” per “Project”. To begin with we might see large variation from project to project. Focussing on this aspect could lead to stabilisation. After a while, there might be optimisation. And if we allow people to re-define the process, ideally there will be a stepwise reduction in the amount of resource required per project.*
This brings us back to the beginning: the purpose of KPIs is to provide insights that we can act on. Meaningful KPIs enable us to monitor and re-design today’s systems and processes, so that we can achieve the desired objectives with greater efficiency and effectiveness tomorrow.
*Note: These ideas are neither original nor new. They come from the Quality Management approach used in manufacturing production. The improvement is the result of changing the process by experimentation. Please note that it’s impossible to get improvement simply by setting a target. See: Deming “Out of the Crisis” chapter 11 Common and special causes of improvement.
This post is also available in: Deutsch