Business Model + Analytics. We hear lots about business models (especially on this site!) and lots about analytics in the business world. But what do Business Model Analytics themselves look like?
The below post is a very simplistic, illustrative example of how the topic works conceptually. It is meant for educational purposes and is not specific to any business or brand in-particular.
The frequently discussed Business Model Canvas (BMC) was a conceptual framework designed by Alex Osterwalder et al. to map out business models, analyze them, and blueprint strategies for new models.
From there, we normally start looking at revenue, cost base, profitability, etc. as metrics to assess a business. But as was discussed recently in ‘Loyalty Metrics,’ there are certain metrics beneath the hood that can give entrepreneurs, executives, and investors a better picture of what to expect going forward.
On a purely observational basis, looking at loyalty metrics of the top 20% of customers can – in many cases – give indications as to the underlying 80% of the volume within the business. The nature of said volume could be different depending on the type of business – transactional value, orders made, repeat purchases, etc. – but based on Power Laws and the Pareto Principle, there is likely some kind of pattern there waiting to be discerned.
This is where the concept of Business Model Analytics begins to come into play. If there are data points that sit beneath the metrics, that data can be analyzed and repurposed to drive the metrics that ultimately affect top-line revenue and bottom-line profits.
Analytics > Metrics > Profitability
Let’s break down certain categories to show some examples of what this looks like in practice. Keeping in mind that every business model has different levers, so not every example below will apply equally (if at all) to each business model.
>a typical metric for businesses is AOV (average order value)
If you were able to drive up the AOV (average amount spent per order, per year, etc) for a given business (depending on the sector), you would increase revenue. A traditional strategy is to begin to achieve this is to raise prices (through different strategies) or bundle in other products (cross-selling) to increase the average value of transaction.
But the question is, what types of behavioral analytics exist on customers who are buying a given product/service at today’s price?
Many companies will look at metrics such as NPS (Net Promoter Score) to gauge current customer sentiment.
The rise of customer surveys and 3rd party review sites such as Trust Pilot can also be used to glean behavioral data on customers:
>what do they like?
>what do they not like?
But from a statistical standpoint, these are generally all averages. A 4.7 Star review for a business is an average, same with an 8.3 NPS.
The question is if the AOV (ie. average) is let’s say $200, what is the distribution of that value? Does a large % of the upper end of those orders come from a certain number of those customers? What types of behaviors define that customer segment?
> Demographic Factors: age, sex, etc.
> Geographic Factors: proximity to business, geolocation relative to shipping costs, etc.
> Socioeconomic Factors: disposable income, cultural factors, local currency fluctuations, etc.
If theoretically, 20% of the top-performing customers could be linked towards some kind of underlying behavioral analytics, this could then, in turn, be used to drive the top-quartile of the distribution and increase AOV without having to rely on the traditional blanket strategy of applying an action (ie. raising the price) to all customers.
>a typical metric for businesses is % of repeat buyers (ie. retention)
In the post on Loyalty Metrics, we talked about different strategies to both qualify and quantify brand loyalty. What we do know is that loyalty can have a significant effect on profit margins for a business. The higher the number of returning customers, the less a company needs to spend on marketing and other similar cost items.
Let’s say that hypothetically a given business had 30% repeat customers in the last quarter. This is a metric that can be tracked and analyzed on a quarterly basis.
But of that 30% of customers, it is unlikely that the distribution is equal among all of them for factors such as:
>how long have they been customers? (distribution)
Imagine that 75% of repeat buyers had been customers for more than two years. This type of data could inform a lot of different types of analysis on loyalty.
>how much have they spent in aggregate? (distribution)
You would expect that X segment who had been there for > 2 years would spend more in aggregate than Y segment who had been there for < 2 years. But what if there were outliers within Y segment. Could that be explained by some kind of strategy that had been applied or other important factor?
>do they take advantage of discounting/sale periods?
Many times loyalty schemes use points and other ‘carrots’ to incentive behaviors. In this case, if a company creates a ‘loyalty discount,’ then they expect repeat buyers to use it. But in many cases, repeat buyers like a brand because it gives them what they are looking for. They are not always price-sensitive. Being able to identify what factors lead to repeat buying behaviors is critical to understanding how to drive loyalty up.
Attribution of certain analytics to certain retention behaviors could open the door to fine-tuning the marketing strategy and loyalty incentives towards customers who are both repeat buyers and those who aren’t. Knowing what customer analytics matter beneath the surface can help increase retention without having to reinvent the wheel.
>a typical metric for businesses is CAC (Customer Acquisition Cost)
More and more companies are tracking CAC now (measuring aggregate ‘demand generation expenses‘ and dividing it by the number of ‘new customers‘ over a certain time period). Investors heavily scrutinize CAC as a way to determine the outlook for future profitability and also use it as a way to gauge the efficacy of the management’s strategy. If a period of strong revenue growth is directly correlated to a rise in marketing expenses (from a P&L perspective), that doesn’t mean the business is in the clear. We have seen many venture-backed businesses fail for precisely this reason.
Without getting deep into a discussion on Unit Economics, CAC is a direct result of the marketing channel that the customers originate from. CAC also typically gets more expensive as a market matures, as more competition equals more companies vying for customers.
In the modern age of digital media, Facebook, Google and other Big Tech companies dominate the landscape for marketing dollars; however, there are several channels other channels that can also be used to acquire customers. An example distribution is:
> Paid Digital Marketing (PPC, CPA, etc): Google, Facebook, Instagram, Pinterest, etc.
> Affiliate Marketing (via 3rd party affiliates): 3rd party advertisers that take a % of each sale for the business
> SEO (Search Engine Optimization): high organic keyword ranking for terms that lead to conversion in X industry
> Referral (from current customers): word-of-mouth referrals from online and offline channels
> Walk Ins (for physical retailers): people walk by a given business and enter it for practical reasons (ie. a haircut)
Naturally, there are other channels, but that covers the majority. It is unlikely that most businesses gain customers from each channel. For most, it is likely that the lion share of customers come from 1 or 2 channels depending on the business.
Most businesses don’t go a level deeper and measure analytics associated with the conversion from those channels outside of CAC (and any other data delivered via the platforms that they use). They do not do competitive keyword research (for example) to see how their competitors use paid channels and how they rank on organic keywords. They do not consider the way that channels can affect outcomes. Here are some examples:
> many companies advertise on Instagram/Facebook/Google because their competitors do, but not have any concept of how their competitors advertise. This leads to bloated CAC as a result of inefficient targeting/budgeting, yet it still brings in customers. Result: High Marketing Expenses relative to Revenue
> many companies do not have any keyword research for their industries, nor from their competitors. As a result, their SEO channel is simply a ‘random path‘ based on hundreds of actions over the years that lead them to randomly rank on some keywords without knowing why. As the CAC on SEO is usually much better than Paid Marketing, many opportunities are missed. Result: Huge loss of low CAC customers relative to competitors
> many companies do not understand the mechanics of referral marketing. They don’t know how much it costs to bring in a new customer, so they tell their current customers to ‘tell your friends.’ Likely, the CAC on any Paid Marketing channel is much higher than the Referral channel, meaning throwing in a small incentive to drive the Referral channel is worth the effort. Yet no action is taken. Result: Referral channel is dormant relative to what it could become.
Further examples could be given for the other channels; however, the overarching point is that having deep data on specific channels is paramount to optimizing them for CAC. A lower CAC means more profitability and historically a better valuation in the market. Channel analytics can be built in relation to each relevant marketing channel in order to better understand how to allocate budgets and target customers on a short, medium, and long-term basis.
HBR Case Study – Flourishers
A 2015 case study in Harvard Business Review (HBR) illustrates some good data in relation to the Business Model Analytics discussed above. Despite 2015 seeming to be decades ago in the modern digital age, the basic concepts are articulated in the same way, showing that the underlying principals still apply regardless of the time frame.
The specific example draws on a national fashion retailer that was struggling with several problems, including “same-store sales were stagnating, and promotional pricing was shrinking margins.” We can imagine that if this was a problem for these types of businesses in 2015 what it would be like now!
Where did this ‘national fashion retailer’ focus its efforts? On cutting costs of course …
The authors worked with the fashion chain for a period of two years. The strategy was to segment and quantify the value of what they called the ‘Flourishers’ segment.
In hindsight, the results on the top and bottom line, over the mid-term, were probably the difference between life and death for this particular retailer. But even without hindsight, we can see that the strategic insights gleaned from this type of analysis were enough to increase revenues, preserve margins, and refine customer acquisition strategies.
That is why this type of ‘analysis’ can pay off for even small and medium businesses (SMBs). The classification of ‘customer segments’ relative to certain Business Model Analytics could be the difference between life and death given the current economic crisis we are living through. These above-mentioned analytics – behavioral, retention, channel – may lead to strategic insights that allow a company to repurpose existing resources and reallocate budgets to boost revenues and preserve margins, and come out the other side of the crisis.
Recent Example of Business Model Analytics – Airbnb IPO