One of the key things my clients almost always ask me is what to measure? How do you make sure you get the right picture, or what we in industry-speak call 'Key Performance Indicators' (KPIs). And it's such an important question.
It's important to note though that what KPIs you focus on depends entirely on your specific business. No business should measure the same things. It all depends what you do as a brand, what products you make, your business and pricing models, where your income is coming from (direct, indirect, advertising), your distribution networks and so forth. And each approach has a different set of KPIs.
Overall, though, any good KPI framework must cover a number of overall metrics, regardless of how you achieve them. These are metrics for:
Note: BFF is short for 'Best Friends Forever', in case you didn't already know.
It's such a simple but wonderful concept which I also use in this article.
The lessons learned from running a businessBefore we start though, there is one major problem facing the industry when it comes to analytics, and it is that too many people are distanced from the actual process of running a business.
One of the key things you learn as an owner of a business is that all metrics are useless unless they can help you get real money into a real bank account. It's the first thing you learn as a business owner. Fancy terms like engagement, awareness, recall rates, traffic, pageviews, time on site, assisted conversions, and so on, are completely meaningless... unless you also get real money.
As Bob Hoffman recently wrote about chasing false goals:
There will come a day when you have to make an important presentation.
There will be a roomful of high ranking clients. You will be presenting the results of an expensive advertising program. You will have charts and graphs and Powerpoint slides and video clips and all manner of stats and data.
You will describe to the attendees how awareness numbers have risen sharply; you will show a chart that demonstrates increasing purchase intent; you will explain how the media buy delivered 8% lower cost-per-thousand than the plan anticipated; you will explain the impressive social media metrics that were achieved, and that brand attributes have risen sharply.
Your contact person at the client organization will be sitting proudly watching you deliver the good news.
But there will be a problem. The problem will be that there will be a smart person in the room. And the smart person will ask one question: What happened to sales?
You will dance and sing and whistle and do ventriloquism. But if the answer is anything other than sales increased, you're dead.
Note: Bob was the founder and CEO of Hoffman/Lewis advertising, one of the West's largest independent advertising agencies, until he retired in 2013.
This is something we all encounter all the time. Part of the problem is how most companies are organized, in that sales and marketing are kept separate and distanced from each other.
For instance, the fashion company I once worked for would not allow marketing to even see the sales figures. Here we had a group of highly talented marketing people, working completely blind... as a corporate policy. How would you know if your metrics mean anything, unless you can see how they influence sales?
The other problem is that people's salaries are rarely linked to how the company performs. We see this, for instance, with the media industry. Have you noticed, for instance, that most of the people arguing against subscriptions for online newspapers are all people with a steady monthly income?
And this is one of the key things you learn when you start your own business. As a business owner, there is only one metric that really matters. How much money do you actually have in your bank account. Is it enough to pay your bills? Is it enough to invest in the future? And are your cash reserves increasing or decreasing?
No matter what else you measure, you have to start with this. Money is the one and only metric that must be at the very top, every time!
However, while money is the first metric of all, we all know that actually earning money is never the result of a simple action. Sure, sometimes you can earn money by simply choosing the right location, like when you open up a gas station at the right distance from other gas stations on a highway. In that case, the simple location of your business will give you the sales you need.
But for most brands, earning money is almost always the result of your ability to influence people over time, and build up enough momentum to generate a sale. It is then something that you follow-up on by further increasing your influence, causing people to become loyal to your brand, which encourages people to come back again and again.
This is why any good KPI framework must include these five elements: Your primary metric is sales, for which you judge the true impact of all your other metrics. Followed by growth, in order to make sure you have the right perspective. Followed by loyalty, to make sure you are achieving long term results. Followed by behaviors, helping you understand your customers and refine your products. Followed by trends, to make sure you are preparing for the future.
The last one is particularly important in industries that are being disrupted, like the newspaper and magazine industries (but also for marketing using those channels). One thing people often forget about trends is that the old world is often still growing, while the new trends are forming.
The cable TV industry, for instance, is still growing, even though we all know that in the future, the old concept of not being able to choose when and what to see will become obsolete.
It's not enough to just measure sales and growth. You also have to put that into perspective as to where the future is heading. And it's the job of your key performance indicators to make sure you get that perspective.
Look outside your internal metricsGaining perspective illustrates another element of KPIs, which is that it's often not enough to just use internal data. Especially not when tracking trends.
Imagine you are a brand that has been spending most of your marketing budget on print and TV. Both have been good channels for a long time, but what about the future?
If you only look at your internal metrics, you might not be able to spot the real patterns. You might be missing key data points (like what your demographics are really like), or things might seem to be 'business as usual'.
So, you look outside your internal metrics and find reports like the one below from Ofcom's "Adults' Media Use and Attitudes Report 2014" (PDF). They asked people which media activity they would miss the most, and the answer showed a dramatic difference between each age group.
All the younger generations absolutely cannot live without their mobile phones, with TV usage in decline. Literally none of the 16-24 year olds can consider traditional newspapers or magazines a vital part of their lives. And desktop web is standing still.
The older generations, however, spend far more time watching TV, a need which is increasing. And it's the same with traditional newspapers. They would readily skip smartphones, with 65+ year olds preferring traditional news over mobile as a whole.
So from a trend perspective we see this generational divide, which I have written about many times before. Not only is there a divide, but it is getting more and more polarized at each end.
This is an important trend, but what is even more important is how this has changed over time. To see this we need to compare it with an earlier study, like the graph below from the same 2010 study:
What you see here is just how quickly people change. The younger generation has been getting increasingly mobile, and more to the point, the 16-24 year olds in 2010 give the same as the responses as the 35-44 year olds in 2014.
In other words, in just four years, the shift has moved 20 years in terms of what each generation is thinking. That is an astounding trend.
Insights like this are as much a key performance indicator as how much money you make or how much traffic you have. It tells you something about the future. If this has happened in just four years. Imagine what will happen in the next four years. How will this influence your brand and your products? How will it impact your marketing strategies? And for publishers, how will this impact your formats and editorial purpose?
Are your internal KPIs aligned with the trend KPIs? And if not, what can you do about it? Looking outside your internal metrics is a critical element to making sure you get the right perspective.
So let's look at four different types of business models. From the media industry, we will look at subscription-based magazines versus advertisement-based magazines, and for brands, retail versus online. This will demonstrate how our KPI frameworks dramatically change their focus depending on each model.
Subscription-based KPIsFor a company based on subscriptions, whether that is a newspaper, a magazine, or a brand selling their goods in some form of membership, its key business objective is to build up momentum.
This is the key element that sets subscription-based business models apart from all other business models. With subscriptions, pretty much all your sales depend on your ability to build up so much value that people decide to commit to you. And the continual existence of your business depends on how well you can keep up that momentum over time.
Obviously, the first metric we need to look for is money (as explained earlier), as in revenue. And, as mentioned by Avinash, we need a best friend (BFF) for this that can help us put this into perspective. The BFF of revenue is profit, because it helps us understand whether the money we make is real or not.
But we also need to look at how that money is earned, which in this case is via subscriptions. We need to track the total number of subscribers. What is the BFF for total subscribers? It's uniques... as in what is the rate of unique people who are also subscribers?
And obviously, if your business is based on multiple subscription plans, you need to track this for each one separately.
These very simple overall numbers tell you something about the current health of your business. But we cannot leave it there. We also need to know how that affects us over time. To do this, another set of KPIs is the number of new subscribers and your churn rate (how many you lose in comparison). By the way, churn rate is the BFF of new subscribers. That's how it works.
Here we encounter the first mistake that companies often make. Most companies measure this using the standard analytics way of looking it per month or per week. But unless you are a monthly magazine that makes very little sense. Instead, you should always measure this in relation to your specific product or campaign cycles.
Now we get to the really fun part. Your next KPIs are "Subscriber-assisted traffic and conversions" and "Marketing-assisted traffic and conversions score". Both are compound metrics that combine several individual metrics into a bigger one that gives you the big picture. And both are growth metrics.
Subscriber-assisted traffic is traffic or customers that you are gained entirely through your subscribers. It's the effect of the loyalty of your subscribers when they share, talk about and recommend you to people around them.
And it doesn't matter how this is created, whether it is via sharing on a social channel, people sending an email to a friend, or other channels. What's important here is to learn how much growth your subscribers are responsible for.
The challenge though, is that you can't measure this by default. This is something you have to build into the very structure of your site or app. On this site, for instance, I have made this part of the URL itself. If you look at the URL of this page, you will notice a weird string of characters after the URL. This is a unique subscriber code that uniquely identifies who this article belongs to. And it's different for each article and for each subscriber.
This allows me to not only make sharing a million times better (since subscribers can now share Plus content), but it also gives me the Subscriber-assisted traffic and conversions metrics. I can simply segment my data on whether or not this specific line of characters are present or not.
In comparison, marketing-assisted traffic is all the traffic that isn't caused by your subscribers. It's the traffic you have to do something extra to get. Either by optimizing your SEO, advertising to get people's attention, or all the other marketing activities that you do.
And remember, what we are talking about here are the KPIs. We don't really care about the details. We want to learn the big picture stuff, which in this case is two things. How much of our traffic is caused by our current customers sharing and recommending what we do, and how much we are attracting through our own efforts?
So what is the BFF metric to compare that against? It's "cost per acquisition". As in how much money and time you have to put into each one
For instance, you might learn that your own efforts generate most of your conversions. But it's at so high a cost that the lower subscriber-based conversion actually generates a better ROI in the long run.
Now it's time to look at our loyalty metrics, or for subscriber-based business models, our momentum.
Loyalty, or momentum, is the key indicator that tells us something about how people use the site. The KPIs here are:
- Unique traffic rates
- Return rates
- Number of pages seen (as a rate per person)
- Read rates
- Engagement rates (do people do something after they have seen the article, and is that different if they have also read it? Surprisingly, the answer to this is sometimes no.
And what you want to be looking for are signs that people feel a greater and greater need for what you have to offer them.
Remember, it's not about the individual metric. A person visiting two times per month, but each time seeing 10 pages, is often less valuable than a person visiting 15 times per month, but only seeing one page each time.
Why, you ask? It's because the person only visiting twice a month is at a much greater risk of forgetting to come back than a person who visits every other day (even if the total pageviews are lower).
Remember, subscription based business models are all about momentum. Traffic without momentum is extremely dangerous to your future.
Finally, we have the article-based behavioral metrics, namely article read-rates, discovery-rates and share-rates for each article (or product) that you publish.
What is the BFF for these? Well, it's the loyalty rates above. You compare the performance of each article to the loyalty rate of your audience, and you compare the loyalty of audience with sales. What you will often find is that these are rarely aligned (which is one of the biggest problems for media companies today).
To summarize. This would be what your KPI dashboard should include for a subscription-based company:
And you then compare this to your KPI trends from the market as a whole, as I wrote about earlier.
Now let's move on to advertisement-based companies.
Advertising-based KPIsThe main difference with advertising-based KPIs and subscription-based KPIs is that the key element is no longer about momentum, but rather about what actions people take. This is exactly why so many media companies struggle trying to combine the two.
Like before, we start with revenue and its BFF profit.
But then, instead of measuring the number of subscribers, we measure the number of ad spots filled over any given time. Its BFF is your ad-rates, and how they are evolving. Obviously, having 90% ad spots filled but at only 60% of the rate is not a good sign.
The next vital KPI to measure is how you sell your ads. This will obviously vary from one company to the next, as each one does it slightly differently. For instance, the Financial Times have just partnered with Chartbeat to sell ad space based not on clicks or views, but on how much time people spend paying attention:
We can report back to a client and say 'we served you a thousand ads, and of those, 500 were seen for one second, 250 were seen for 10 seconds and 250 were seen for 30 seconds. The next obvious step is to sell blocks of time.
We can sell a thousand hours of exposure to a chief executive audience in Germany, for example, or we can give clients 500 hours of exposure to finance directors in Belgium. That currency has a lot of merit.
In this case, their KPI is time a person is paying attention (which is a more advanced form of read-rate). For other companies, this KPI might be clicks, exposure, or other actions that your audience performs.
It all depends on what metric you are selling you ads spots with. This metric is in many ways as important as your revenue metrics. It's the basis for your entire business.
However, this is a bit tricky. It's actually the same problem that we see with native advertising. The question you should ask is, does it actually help the advertisers if people see something for 30 seconds, rather than just 15 seconds? Or is time even the vital factor that causes people to buy the advertised product?
Today nobody really knows, because we lack comprehensive studies to uncover it. And there is no way you can measure this internally. This can only be done via outside studies and it's linked to 'trend KPIs.
Next we want to look at how hard or how easy it is for you to sell your ad spots. We do this by looking at how many of the ad spots you sold organically, and how many you had to get through the efforts of your sales people.
In other words, how many ad spots were sold because brands contacted you on their own? And how many were caused by your sales people calling customers to persuade them to buy?
Obviously, the more ad spots you sold organically, the better.
The BFF metric for this one is the loyalty of your advertisers. What is their return rate? And how often do they place new ads? This is a vital metric in judging your advertising effectiveness. Brands that come back often only do so if they believe that the previous ad performed well.
Of course, it isn't just about advertising, you also need to compare this against your content. You need readers, and you need them to come back, and you need them to do it as often as possible.
So like before, your content KPIs are "Reader-assisted traffic" versus "Marketing-assisted traffic" in relation to your advertising metrics. In other words, how much of your traffic growth is coming from existing readers, against traffic generating efforts of your own? And how does this compare to how much people pay attention to your advertisers?
This is often where you will find that you have a problem. The main reason why ad-rates are dropping is because everyone is turning to the same link-bait tactics to boost their traffic numbers. This turns their audience into mindless zombies who aren't really interested in what they see (including the ads).
And like before, you need to measure article read-rates, discovery-rates and share-ratesagain in relation to your advertising effectiveness.
Remember that it's important to measure this as a 'rate' as well as what you gain in total. An article that goes viral will often have a high total views rate simply because of how it scaled. But the view rates per person might actually be many times lower and so not that valuable for your advertisers.
Case in point. Here is a graph from Monday Note illustrating the revenue from advertising in relation to traffic.
Remember this the next time you hear someone talking about 'what we can learn from Buzzfeed'. Not to mention Upworthy or Distractify (and hundreds of other 'viral' sites) that are performing even worse, per uniques.
The result is a KPI framework for advertising-based models like this:
Brands + retailNow let's move on to brands, specifically a brand that is selling their products in physical stores. The problem with this model is that your point of sale is disconnected from yourpoint of interest. In other words, all the things you do to get a sale happens on channels that cannot be directly connected to the actual sale.
I had this problem every day, when I was the digital media manager for a big fashion company. For instance, of all the people visiting our stores that day, how many times had they visited our website? We didn't know, because we had no way of linking a website visitor to a retail store visitor.
You just don't know.
Same with our print advertising campaigns. How many people who bought a product that day, did so because of the full page ad that was placed in the newspaper earlier that morning?
Again, we had no idea.
Sure, people will then tell you to use discount codes or other mechanisms that would help you connect the dots. And yes, if the situation justified such actions, that would be useful. But remember, asking people to give a code means adding an extra step, which sometimes has the unfortunate side effect of lowering sales. It creates a form of friction.
It gets even worse when you don't own your own stores. Now you can't even track your sales, because you have no idea how many products the stores actually sold. You only know that three months ago, one store bought 12 boxes. But you have no idea if they have sold any of the products yet, or when they sell them.
This was the problem I had every day. Thus, you can't really look at the specific numbers. Instead, you need to look at the patterns. And you need to look more at the overall brand metrics.
Like before, we start with revenue and its BFF profit. These are the key indicators no matter what. But then enhance those by looking at three other metrics, namely the number of transactions, average order value (AOV) and number of products sold per order.
These five metrics define our overall patterns that everything else has to be matched against.
You then need to look at your other metrics, specifically those defining the value of your brand, the activity around your channels, and (if possible) the aggregated result of the specific outcomes of your individual campaigns.
You determine the value of your brand by looking at things like:
- Brand awareness and recognition (how many people know about your brand?)
- Market penetration
- Brand reputation (the kind you learn by doing sentiment and mood analysis)
- Your consumer satisfaction score (especially in relation to the rest of the market).
- Your website and blog traffic
- Social mentions
- Number of people talking and engaging with you across not your own channels but across the internet as a whole.
Mind you, it's rarely on a one to one scale. If your website is only responsible for 5% of your sales, doubling your traffic won't also double your revenue. It will merely increase it by a few percent. This is why we need to look at this in terms of patterns and not absolute numbers.
It doesn't really matter what the numbers are specifically. You just want to know that your sales numbers followed the same patterns as the activity that happens around your brands.
The final element, "the aggregated result of the specific outcomes of your individual campaigns", is a bit more tricky. The idea is that you look at the details of each thing you do, and you calculate the sales uplift that it created. And then combine all these numbers into a single aggregated KPI that tells you about how your campaign activity is doing as a whole.
If you are running a full-size ad in a local newspaper featuring 4 specific products, you look at your sales before and after the campaign and measure the difference. This gives you a sales uplift rate. So, if you experienced a 20% increase in sales, that is a sales uplift rate of 0.2.
And when you combine this with all your other campaigns, you have a good idea if things are going the right way or not.
The problem, of course, is that you often don't have this data. You can only do this if you own your own stores and if your point-of-sale system is tied into an analytics system on a product level.
But if you can do this, it's a very powerful KPI to have.
Brands + online shopWhat if you are selling your product exclusively online instead? What KPIs would you look at then? The wonderful thing about online shops is that we have much better data. With retail shops we looked at mostly vague data to see if we could spot a pattern, with online shops we can narrow in on specific and much accurate insights.
We start off with the five key performance indicators as before:
- Number of transactions
- Average order value
- Number of products sold per order.
- Number of customers
- Number of orders per customer (e.g. per year).
This part is important because one of the things that kills most businesses is when their customers stop renewing themselves. We see this all the time. Look at Kodak, Blockbusters, Blackberry, and hundreds of other companies.
And it's not just for your brand as a whole, it's also critical to measure when you are launching something new. If you launch a new product and people keep buying the old one, you've got a problem. Because if you can't renew your market you are in a race toward oblivion.
Another area of growth is the same as what we have from subscriber-based business models. In order to understand how your growth is happening you need to look at yourcustomer-assisted sales versus your marketing-assisted sales.
In order words, how much of your sales is occurring through sharing and recommendation by your happy and loyal customers, and how much of it do you have to attract the hard way through expensive marketing campaigns?
The next set of KPIs are those you track in order to learn about your customers behavior and loyalty. What is the rate of sale per customer? As in, how many times per year does each one of your customers complete a purchase?
And how many times do they visit your web shop? How many products are they looking at? How many products are they adding to their carts? What is the shopping cart abandonment rate?
As with subscription based models, where you determine the value of each story by looking at things like read rates, a webshop determines the value of each product based on how people explore each one.
However, there is one key thing to remember. Each metric, in itself, is largely useless. Take shopping cart abandonment rates and consider these two people (for the same year):
One customer is a high-frequency customer who often visits your site, but also often leaves without buying anything. He has a terrible shopping cart abandonment rate and a an equally terrible conversion rate... but he is worth $2,750 to you.
The other customer is a low frequency customer. He has an extremely good shopping cart abandonment rate and an excellent conversion rate. Almost every visit results in a sale. But over the past year, he has only purchased products worth $238, or less than 10% of the first person's sales.
So tell me, what is a good shopping cart abandonment rate? You can't answer that, because it always depends on your overall sales.
And this is why having a BFF metric is so important. You need another metric to compare it against in order to get the right perspective. In terms of webshops, the BFFs for your loyalty and behavioral metrics are total sales, number of transactions and number of products sold.
And of course, never forget about the data that you cannot see. For most brands, this data represents more than 60% of your overall sales. Read more about this in "The Mystery of Our Social Traffic".
Four different KPI modelsWe have now explored four very different business models, and you have seen how each need different KPIs, and that is the whole point. Your KPIs are unique to your business, yet they are all based on the same overall concept.
Brands, of course, often mix retail and web shop, with an increasing need to think more about how they tell their stories, which is very similar to a subscription-based model.
So your challenge is to match your specific situation with the right KPIs, and then look at the trends (something that most companies ignore when looking at the KPIs).
A company like Blockbuster experienced positive growth for many years while it was completely ignoring the emerging digital world. They knew the trends were there. They talked about them, and wrote that it was a 'risk' in their annual report. And yet, they didn't act on it.
There are many similar examples like this in every industry. For instance, in the fashion industry, the introduction of microfibers completely changed how you would make sports and outdoor gear... and yet, most fashion companies were still focusing on design.
Here is an example from Peak Performance (just ignore the oh-so-dramatic voice-over):
Paying attention to the trends and identifying what creates superior or disruptive products is a critical element of your KPIs. In fact, external trends is your BFF metric that you compare with your internal trends. And you have a lot of work to do if the two don't align.