Rethinking Referral Marketing (Part 2)

As I started to think about the new customer acquisition problem, my first answers were:

  • Identify who your best customers are
  • Decode the genome of these best customers
  • Use the Best Customer Genome (BCG) to run lookalike campaigns for new customers

I should mention another point in passing. A brand’s best customers could also be persuaded to get others like them – given that each of us has a wide social network, and we do tend to advocate brands we like to our family and friends. Word-of-mouth has a big influence on what we buy.

I never thought of this seriously until the same question got asked in three successive meetings. In fact, one of the CMOs put it in straight-forward terms: “I don’t want to spend a lot running Google/Facebook ads for new customer acquisition. I have X million happy customers. How can I get them to get me more like them?”

What the CMO was alluding to was referral marketing. (It is also called word-of-mouth marketing, but I will stick to referral marketing in this series.) That got me thinking about the problem more deeply. One angle especially struck me – how to persuade the best customers to do referral marketing? The best customers are very likely to know and refer more like them – which means it can become a very powerful positive feedback loop for revenue growth for a brand.

The problem seemed very similar to the loyalty programs that brands run. Most brands have a loyalty program, but few have transitioned to making it into a VRM program which creates a differentiated experience for their best customers. Referral programs also seemed to be stuck in a similar time wrap – if they existed. A free ride here, a few points there, and that’s it.

Were brands missing out on a big opportunity by not combining the two ideas of best customers and referral marketing? If so, what could they do?

Tomorrow: Rethinking Referral Marketing (Part 3)

Rethinking Referral Marketing (Part 1)

The past few months have given me a lot of time to think deeply about marketing. Starting with the ideas of Velvet Rope Marketing (VRM), I have had dozens of conversations with CMOs and digital heads of brands in India, SE Asia, Middle East and Africa. The basic problem statements are the same: how to get more from the existing customer base and how to optimise new customer acquisition.

In this context, the VRM pitch works well:

  • All customers are not equal; some have more customer lifetime value (CLV) than others
  • Identify your best customers – say, the top 20% — based on CLV
  • Create a differentiated experience for these customers to maximise revenue and retention
  • Among the rest, use individual customer genomes to identify who could be future best customers
  • Also use the characteristics of best customers to acquire new customers who look like them

On paper, this looks simple. But it is quite surprising how few companies actually do this in a disciplined manner.

I have identified eight problems that handicap the use of VRM:

  • Limited data: Brands are not collecting customer data at every touchpoint
  • No CRM / CDP: The collected data is not centralised into a single place to provide a unified customer view
  • Data not being analysed: If the data is there, then the analysis is either not being done or is very limited
  • No internal team / skillsets: The right talent is missing for doing the analytics and then acting on it
  • Incorrect CLV calculations: Too often, the CLV calculations are done in a very simplistic manner (for example, basing it on the average value of past transactions)
  • No tech platform: Even if all the above is done, there is no tech platform to help implement the use cases
  • Adtech-Martech silos: The acquisition process is not streamlined because there is no exchange of data between the acquisition and engagement teams
  • Cannot measure RoI: Finally, it is hard to get the green light for VRM because the CMO is unable to show the actual uplift in revenue

In my presentation, I address each of these challenges and show the solution to every identified problem. These series of essays discuss the VRM roadmap in depth:

And as it happens so often, one idea leads to another. While VRM does well in addressing the needs of the best customers, a question that came up repeatedly in my meetings was about how to optimise new customer acquisition.

Tomorrow: Rethinking Referral Marketing (Part 2)

Podcast on Velvet Rope Marketing

I recorded a podcast with Dennis Dayman as part of the “Email Unplugged” series driven by our US team.

We discuss my early entrepreneurial journey, and then go deeper into VRM:

  • What is Velvet Rope Marketing all about?
  • How can businesses implement VRM during a crisis?
  • How to identify your best customers and communicate with them effectively
  • How to cut down marketing costs with VRM
  • The importance of full-stack automation platform to leverage VRM
  • How VRM can help you acquire new customers
  • Special tips to optimise your email campaigns with VRM

Announcing: hippoBrain

Along with Jaimit Doshi, I have launched a new show — hippoBrain. As we put it, it is about “hippo-sized conversations with hippo-sized brains.” The focus will be largely around data, tech and the consumer.

The idea came a couple months ago. While there is a lot of content out there, what I find missing is deep conversations with people — to really understand their mental models and how they think. In today’s attention-deficit world, everyone is looking for that 30-second sound byte. But what is missing is the context around that. Why does the person think in that way? That’s what we want to do with hippoBrain — give an insight into how the guest on our show think.

We have launched with 3 amazing guests: Vivek Bhargava (Denstu;  ad-tech authority), Kartik Jain (DBS, ex-HDFC Bank; marketing maestro) and Anindya Ghose (NYU Stern; mobility and AI guru). We will have one new hippoBrain episode every Friday.


The One Number To Predict Revenue (Part 10)

What can you do as a business leader? Here are the 7 steps to getting started with the ideas and apply the ideas of NPR in your business:

  1. The starting point is transaction data. Collate customer-wise transaction data for the past few years.
  2. Calculate Customer Lifetime Value (CLV) for these customers. Remember that CLV mentioned here is predictive and forward-looking.
  3. Plot the CLV-Customers chart and see the power law at work
  4. Based on this, you can calculate the Predicted Revenue – the aggregate CLV for all present customers
  5. By factoring in loss of revenues due to expected churn and additional revenue via new customer acquisition, you can get the Net Predicted Revenue (NPR)
  6. The next step is to apply the ideas of Velvet Rope Marketing (VRM) to grow revenue from the Best Customers, get some of the Rest and Test Customers to become the Best, and acquire Next Best Customers who are likely to have a higher CLV (see graphic below)

  1. Set up a process by which the CLV and NPR calculations are automated and available periodically to all decision-makers.

Through these actions, you can now view in near real-time the impact of the marketing activities on a forward-looking basis through the lens of customer actions. NPR should thus become the single business number to discuss strategy and drive new initiatives.

These are ideas which can be applied to every business in every industry. By calculating NPR across time, one can see the impact of customer-centric initiatives. NPR thus allows vendors to become partners because the impact of each action can be measured, and thus gainshare models can be applied to customer acquisition, retention and development. For the first time, business leaders can get a measurable marker for future profit growth. The CLV-VRM-NPR triad is the path for every business leader to rise to the most important position in today’s challenging times – becoming the CPO (Chief Profitability Officer).

The One Number To Predict Revenue (Part 9)

This single chart brings together the ideas of Velvet Rope Marketing and Net Predicted Revenue.

Once the CLV for all customers has been calculated, we get the Predicted Revenue for all present customers. This is indicated by the green and yellow areas under the curve. (Best Customers are the top 20% customers with a CLV greater than CLV0, while the middle 30% customers have been labelled Rest and the bottom 50% as Test.)

Marketing’s objective is simple: grow the area under the curve. This is what VRM does and NPR measures.

There are 5 actions that marketing needs to do:

  1. Make CLV from the Best happen: this is the core of VRM – how to provide amazing experiences to the Best Customers so they do not churn. At a minimum, the projected CLV is what must be realised.
  2. Cross-sell/Upsell to grow CLV from the Best: this is about increasing the green area by pushing the curve higher. This is where a good understanding of the customer genome combined with a martech platform to implement VRM can help create what we can think of as “100% customers” – where a brand monopolises the complete spending of the customer in a category.
  3. Nudge some from Rest and Test to Best: this can be done by analysing the Best Customer Genome, identifying the next best action for each customer, and creating the future pipeline of new Best Customers to replace some who will inevitably churn.
  4. Acquire right Next: this is about using information about Best Customers to acquire lookalikes so that the wastage that is ever-present in new customer acquisition can be reduced
  5. Help them become Best faster: this is about providing a super onboarding process for likely Best Customers to ensure they can cross the CLV0 spending threshold sooner than later

Tomorrow: The One Number To Predict Revenue (Part 10)

The One Number To Predict Revenue (Part 8)

To better understand how NPR can be applied, a business can start by charting the CLV for each customer.

Next, sort the customers by CLV from high to low on the X-axis.

And what do we see? A curve which looks like the power law!

We can then segment the customers into Best-Rest-Test.

Marketers and business owners can then get to work on the customers to grow the CLV. In the graph below, the brown area shows how the predictive CLV changes for some customers leading to an increase in the area (which is the predicted revenue).

In this case, what has not been factored in is churn and new customer acquisition. Also, the customer acquisition costs (CAC) have been ignored.

These charts show how the CLV calculation can help in predicting revenue. Business leaders can now get a forward-looking impact of their present-day customer-centric initiatives. Given that costs are more directly under the control of the leaders, Net Predicted Revenue (NPR) is the one number that can thus be used to predict revenue – and perhaps profits.

Tomorrow: The One Number To Predict Revenue (Part 9)

The One Number To Predict Revenue (Part 7)

NPR as we have discussed is a bottom-up way of predicting future revenues. It starts with the actions of a single customer. Will that customer stay? Will that customer spend? Will that customer churn? Past actions as computed through an RFM (recency frequency monetary value) grid help compute the probabilities for what a single customer will do.

What is important here is the focus on a single customer. There is an equivalent idea in economics that is focused on a single individual as the unit of analysis – the concept of “methodological individualism.” Here is an explanation from The Encyclopedia of Libertarianism:

According to Ludwig von Mises, methodological individualism views “all actions [as] performed by individuals”—or, in the words of Karl Popper, that social phenomena “should always be understood as resulting from the decisions, actions, attitudes, etc., of human individuals, and that we should never be satisfied by an explanation in terms of so-called ‘collectives’ (states, nations, races, etc.).”

So, even as we think about customers as the collective, it is the single customer that should serve as the basis of analysis and action. The idea of CLV assigns a future revenue value to each individual customer. It is a bottom-up model, rather than just top-down extrapolation. It is forward-looking and predictive, rather than taking a number from the past.

In the real world, customers are masters of their own preferences. As such, they can continue with a brand or they can switch to another brand to get their job done. One brand’s churn is another brand’s acquisition. By starting with predicted revenue and layering churn and acquisition, it is possible to arrive at Net Predicted Revenue (NPR).

This becomes the most important metric for a marketer – change in NPR over a period of time will help provide early indications of the future health of a business. It is a lead indicator, a crystal ball into the future. NPR is the single magic number that CEOs who think profitability (rather than chase valuations) should be focused on.

Tomorrow: The One Number To Predict Revenue (Part 8)

The One Number To Predict Revenue (Part 6)

Customer lifetime value (CLV) is important because all customers are not equal. In fact, most businesses will find that there is a power law with CLVs. As explained by Wikipedia, a power law is “a functional relationship between two quantities, where a relative change in one quantity results in a proportional relative change in the other quantity, independent of the initial size of those quantities: one quantity varies as a power of another. For instance, considering the area of a square in terms of the length of its side, if the length is doubled, the area is multiplied by a factor of four.” This is how a power law curve looks:

It is not surprising that CLV curves look very similar! This is similar to the 80-20 rule, where 80% of value comes from 20% of the base. (The specifics may vary – it could be 90-10 or 60-20.) For most businesses, a small number of customers account for disproportionate value (revenue and profits).

Power laws are common in product sales. A few “hits” account for the bulk of the sales, followed by the long tail, as shown in this graph below from TechnoLlama:

There is an excellent collection of charts in a post by Michael Tauberg which shows power laws all around us:

Tauberg explains power laws simply: “Basically, power law is like a forest. There are tall trees which soak up the sun and grow to be enormous. Then there are all the shrubs on the forest floor.”

Below is a real demonstration of the power law from one of our customers – the hand-drawn green curve shows the power law at work.

The lesson for business leaders is to recognise the presence of power laws in their revenue charts. Calculating forward-looking and predictive CLV brings this to life. It also leads to another conclusion: businesses and marketers need to take exceptional care for customers who are at the high-end of the CLV charts. Focusing on these customers is what Velvet Rope Marketing is about.

In this context, what Net Predicted Revenue (NPR) does is provide a glimpse into the future. It shows what revenues a business can generate going forward from all its customers. The good thing about NPR is that it is calculated one customer at a time. Making a single customer the unit of analysis takes us to one of the most important ideas from the world of economics.

Tomorrow: The One Number To Predict Revenue (Part 7)

The One Number To Predict Revenue (Part 5)

Before we discuss the power of Net Predicted Revenue (NPR) as marketing’s magic number and the one number which according to me ‘solves marketing’, let’s start with the basics.

A company has customers. What it knows for sure are its past transactions with these customers.

Aggregate all these transactions in a specific period and you have the revenue for a day, month, quarter or year.

Just by looking at these numbers it is not easy to predict the future. One could extrapolate based on growth. So, if revenues have grown by 20% for each of the past three quarters, one could expect future revenues to also grow by 20%. But that does not tell the marketer who will spend how much and which customers are at risk of churning. This is where the concept of customer lifetime value (CLV) comes in. It is the present value of the future revenues from a customer. (The calculation is akin to the discounted cash flow model that companies use to calculate current valuation.)

Once we are able to calculate the CLV for each customer, we can get an aggregate view of future revenue from all customers – think of this as the Predicted Revenue. This is at a single moment in time. It is simply the area under the CLV curve, where the Y-axis has the CLV numbers, and the X-axis sorts all the customers from the highest CLV to the lowest.

So, by using CLV for each customer, we are able to calculate the forward-looking revenue for the business. What we then have to factor in is that some customers will be lost (churn) and new customers will come in (acquisition). Taking both these changes leads us to Net Predicted Revenue (NPR).

Tomorrow: The One Number To Predict Revenue (Part 6)