How FMCG companies can use E-commerce to broad-base their portfolio

Dear Finance Gurus on Twitter,

Pranam!

This is FY41 guy. Yes, the same guy who forecast FY41 sales for Nykaa. Over the past few days I have experienced your love and affection, with so many of you calling me Sanjay from Mahabharat, Albert Einstein, and the blind guy who led 175 unseeing anchor investors to invest in the IPO. I only wish I had the powers attributed to me.

First of all, I am so sorry for replying so late. When I am not looking at my crystal ball 20 years into the future, I am bogged down with analysing quarterly results and listening to their conference calls. But since you have entertained me so much, I thought it only fair to respond to your kindness. 

So, a lot of you have been asking me as to where I purchased this crystal ball. As if I'm going to reveal that secret! I am however, going to let you in on the secret of the process. So here goes. 

Why DCF? 

What took me by surprise amongst all the time and attention you have spent on me and my report, is that the discussion was not about whether my projections were too conservative or not conservative enough. It was about why the projection that far out was made at all. To give an analogy, in any debate, if a person questions another person's argument on merits, that is part of the rules of engagement, but if one questions the other's locus standi to make an argument in the first place, where do we go from there? To all those who ask, why do you have an FY41 projection, I say "Sir, have you never done a DCF?"

As I'm sure you're aware, the value of a stock is derived from the discounted cashflows generated over the life of the company. Just think about it - the market cap of HUL is 60 times the profit it generates in one year. Why? It is not because it promises high growth over the next 3 or 5 years, but because it can deliver boring growth for several decades. Even high FCF generating companies such as HUL derive over half their value from the cashflows generated  AFTER FY41. 

We often forget this fundamental law of finance, because we are so used to valuing companies on PE or EV/Ebitda or other such valuation metrics. Indeed, in most cases my target prices are based on PE or EV/Ebitda multiples and not DCF. However, these are only shortcuts or proxies for a DCF valuation. These can be fairly good proxies for companies which have been listed for a while and have a listed peer set and I use these proxies extensively. 

Why DCF for Nykaa?

IMHO, there was no reliable method of valuing Nykaa except for DCF. The business is nascent and profit margin is currently low. PE multiple looks ridiculous. On IPO price FY21 PE is ~850x and even on our FY23 EPS it is over 250x. How do you make sense of these numbers? Is 250x ok or should it have been 150x or 400x? Price to sales is 21.7x FY21 and 11.5x FY23. Is 11.5x reasonable? Should it be 8x or should it be 15x? How does one know? It is not a listed stock, so one doesn't have a historic valuation to rely upon. There are no listed peers in the space that it can be benchmarked to. If and when Purplle or MyGlamm lists, they can use Nykaa's valuation as a benchmark. But for Nykaa, whose valuation should I use as a benchmark?

Some investors have even asked me for Nykaa's valuation in light of Zomato's valuation. I think that's a pretty ridiculous thing to do. The two companies are catering to two completely different needs, the market sizes are different, market structure is different, business models are different. The only thing common is that they both use the internet to market their products/services and fulfil their orders. Would you compare two different businesses just because they are brick and mortar businesses? Then why do you do it just because they are internet businesses? 

You can take Zomato valuation as a starting point or Nykaa, but then you will need to adjust for all these above mentioned factors in the price to revenue multiple (more complication, because these companies book revenue differently - marketplace commission vs net sales. Even GMV doesnt help - there are no cancellations or returns in Zomato, but are part of Nykaa's GMV). How much do you adjust for each of these differences? Should you give a 20% premium to Nykaa for the fact that it is profit making and Zomato isn't? Or should it be 50% premium? What is the logic or basis for either of these numbers? And how much discount for lower market size? How much premium for lower competition? And so on. There is no logic to determine the premium or discount for any one point of difference, and there are multiple points of difference. You might as well be picking the price band directly out of thin air. 

In absence of any historic price data and suitable companies to compare the valuation to, the only option is to ditch the proxies and go to the fundamental way of valuing stocks - DCF. 

Why such detailed forecasts? 

Ok, you might say, do the DCF, but why have such explicit forecasts till FY41? Isn't it a norm to make explicit forecasts for say 3 years, and then just directly forecast a growth rate in FCF for another ~15 years before plugging in the terminal growth? Yes, that indeed is the norm and that is what we do when we do a DCF for established consumer companies like say HUL, because we know what the growth rate is likely to be based on overall FMCG growth rate, which in turn is linked in some way to GDP growth. So if you were to ask me how much could HUL topline post FY25, I'd say 9-10% and you might say no, 12% and another might say, no 7%. 

But that is the extent of disagreement. Nobody is going to say 20%, and nobody is going to say 5% either. However, in case of Nykaa, we are in uncharted territory. Nykaa is currently growing at over 50%. How long can it grow at this pace? When will it slow down? How much will it slow down? It depends on how many new customers they can acquire and how much a customer spends annually. Will the new customers spend the same as the old? If they spend lesser, will the growth in the annual spend of the old customers compensate for this? 

These are all points worth pondering, and if you read my report fully rather than scoffing at one chart without reference, you will see that I have tried to think through these questions - whether you agree with the numbers I have arrived at is a separate discussion, but it is important for each investor to think about these aspects and arrive at his own conclusions. What I provide to my clients is a framework for thinking - which in turn is embodied in the financial model. I have had several institutional investors who have asked for my model so that they can change the forecasts and see how it changes the results. 

Nykaa has 5.6m annual transacting customers. How much can this number go to? That will depend on what their target audience can be. We have defined the target as 35m customers - and there is a logic behind that number. How fast they can reach that number will determine their revenue. Ultimately, the value of having long term forecasts combined with a DCF model is not that it gives you a firm answer as to what the valuation is, but rather what the range of valuation can be based on different assumptions and what is the sensitivity of the valuation to different assumptions. For example, if you were to have a lower number of customers with a higher average order value, or vice versa, with both alternatives giving the same revenue, the profit would be higher in the former alternative. How much does that change the value of the firm? That is worth knowing. And then it is worth thinking if the AOV will increase from here on because of normal growth in spending by each customer, or will it come down because the new customers beyond a certain number will not be as affluent as the current customers. And one needs to think of these questions on a multi year horizon. 

At this stage, I would like to clarify three points

1. A projection is not a prediction. I have heard this from finance gurus on twitter - one does not even know what will happen two quarters down the line, how do you know what will happen 20 years? Simple answer - I don't know what will happen. But since the value of the stock is what discounted FCF the company is likely to generate over its lifetime, I need to have projections which seem reasonable at this point of time. What is reasonable? Reasonable is what consensus deems to be reasonable - the weighted average of buyers' and sellers' assumptions. Should I be worried that these numbers will change over time? Maybe in 2 weeks, let alone 20 years? No, I'm not worried about that. Assumptions change based on company performance, action by competition, government regulation, state of the economy and a multitude of other factors. 

Change in forecasts is one of the most common and most important drivers of change in stock price (the other is change in discount rate, but that is a topic for a separate blog). In absence of change in forecasts, volatility in stocks would come down massively, and stocks would behave like bonds. A projection is not a single number cast in stone. It has a range of probabilities and the single number in our model or report just denotes the probability weighted average number or the most probable number (all this is implicit, not explicit). As company performance, competitive action, regulation, economy etc change, we will change our forecasts. 

Some changes can be anticipated - one could, effectively argue, that although competitive intensity is low right now, it will increase and you should build it into your estimates now, without waiting for it to materialise - and that is a valid point of view. Or one could argue that even if competition comes, Nykaa has moats to protect itself due to so and so factors. A weighted average of these two views amongst all market participants is what will get factored into the valuation. As and when competition comes in and depending on its initial signs of success or failure, the weighted average of the two scenarios will get tilted more towards one side. But refusing to forecasts, because forecasts are subject to change is a weak argument. Because the market demands a bid and ask price daily. 

2. A research report is an advertisement. The fact that Nykaa is a nascent company with a long growth horizon ahead of us is an even more compelling case to have long term projections and a DCF model. The real value however, comes out of tweaking different line items of the model and examining the sensitivity of the changes to the final output such as profit or valuation. Our report contains what we felt were the most likely or reasonable estimates, but it is important to have a bearish and bullish case in addition to our base case. Since our report was already 61 pages long, and there are so many parameters (customers, freqeuncy, AOV, fulfilment cost per order, employee cost, etc) and each of them can have a bear/bull/base case, the permutations are enormous and not suitable to be included in a report. Rather, a research report sent out to our institutional clients advertises the fact that we have done detailed work on the company and that they can reach out to us, have a discussion and use our model to input their own estimates. The process of forecasting and valuing a stock based on DCF is more important than one final number. The data thrown out by the iterations is more valuable than one final outcome.    

3. Stock prices are not always fundamentally driven: I am not so naïve as to believe that stock prices behave exactly as per DCF models. There is a lot of noise in the stock market. There is a lot of volatility. However, by and large, markets are reasonable. And in fact, if stock prices deviate materially and for a long time from fundamentals, is that not something that you would want to know, so that it can be an input into your decision making? To say that stock prices anyways do not behave per DCF and therefore why do it, is robbing yourself of valuable data. You can choose to ignore that data, you can take a call that liquidity or whatever other factor will cause the market to remain disconnected from fundamentals for a while, but it is important to at least have a sense the quantum of that disconnect and whether you are comfortable with it (although, in many cases, liquidity and other technical factors can be actually factored into the DCF via the discount rate, and therefore share prices may not be too disconnected from DCF value in the first place). 

So, in summary

1. I did a DCF because there were no comparables in India which I could use as a benchmark for ratios such as price to sales etc (companies outside India, even if comparable in terms of business are not useful for a valuation benchmark as Indian companies always trade at different multiples due to a multitude of factors).

2. I had explicit forecasts for 20 years for all line items rather than just an FCF growth after 3 years because this was a company at an early stage of evolution and it is difficult to instinctively arrive at a long term growth rate like in the case of FMCG companies without thinking through the individual line items. While this is true for sales, it is even more important for margins as operating leverage plays out over several years and not just 3-4 years.  

3. Forecasts are not cast in stone. They change based on several factors. Change in expectations is the number 1 reason for change in stock prices. Not forecasting because forecasts will change is a weak argument. It is like saying I will not bet on the outcome of a cricket match because the odds will change after after the next ball. 

4. DCF is important because you know what you are paying for. If you do not have a model with long term forecasts and a DCF done, you do not know whether the 53k cr asking price is reasonable or not. I hear ridiculous statements such as "the IPO is expensive at a price to sales of 22x". How do you know that 22x is cheap or expensive? Based on what? But under my method you can say "I think the asking price of 53k cr is expensive, because they would require x number of customers by FY30 which I think is not feasible, and that their AOV needs to go up to y, which is a xx% Cagr over FY21, which will not happen because the new customers will not be affluent enough to spend that much". Or something like that. You know what you are paying for. 

5. You can play around with different line items to see what the business model is most sensitive to. You may realise that a 2% drop in AOV does not offset a 2% increase in customer acquisition (numbers hypothetical to illustrate the point). And you will know by exactly how much shareholder value is being eroded in this trade off. If you see that the company is pursing customer acquisition at the cost of AOV, it will give you an input to your decision making. You can take a view "ok this just a temporary push being made by the company and I'm ok with it" or "no, the company seems to be more interested in just recruiting customers but the sacrifice of AOV is not worth it". Without a DCF model, which factors these line items individually, you have zilch - no input in your decision making. 

6. The point here is not whether I will be proven right or wrong (in terms of the stock call), because that is not the point that you raised - you never expressed a view whether the projections in my report were too high or too low, but rather why the estimate was made in the first place. So in case I do get it wrong, it does nothing to vindicate you. In fact, I would not even have reacted if the problem was that you thought my estimates were too high - because you are a market participant, and you are one of the views which go into the consensus. And consensus estimates of long term growth is what makes the stock price. My response is because of your disdain for me having an estimate in the first place. And I hope I have clarified why you are wrong on this count.  

7. Before trolling someone, consider the possibility that you may have given the topic much less thought than the person whom you are trolling has.