Modeling the Big Disruption
View attachment Market Share Model.xlsI would like to share my disruptive market share model. This model is an extension of the logistic growth model I presented in a tutorial a few days ago. I that model I made the assumption that Tesla would ultimately gain 10% share of the global auto market, but wouldn't it be nice to have a model that anticipated how much market share Tesla could gain under certain assumptions of the competitive intensity of Tesla and the rest of the auto industry? That is exactly what I propose to do. The aim of this model is not so much to predict the future as it is to understand the implications of competitive intensity, to explore sensitivities to the strategic choices of competitors, and to consider strategic options. To that end, I post a working worksheet model for readers to play with and a sensitivity analysis table to summarize outcomes under different strategic choices. Before I discuss the model, let me summarize a few important observations from this model given Tesla's current level of competitive intensity:
1) Tesla should be able to deliver 500,000 cars in 2020 regardless how strong or savvy the competitive response may be. The industry cannot slow Tesla by competing with it.
2) Conventional ICE auto sales will peak between 2022 and 2025 after which conventional autos will go into perpetual decline. This will be a critical moment for traditional automakers.
3) At current growth rates, hybrid vehicles may never gain more than 10% market share and will likely peak between 2028 and 2031. At about the same time,
the entire fleet of conventional hybrid autos will peak and go into perpetual decline. This will be a critical event for the oil industry.
4) Tesla may well reach sales in excess of 21 million vehicles by 2030 and may ultimately gain 84%, 55%, or 25% of the global auto market depending on whether the industry pursues EV growth with low, moderate, or high intensity (30%, 40%, or 50% growth per year). The industry would actually have to sustain an EV growth rate high than what Tesla is committed to to prevent Tesla from capturing more than 12% market share.
These claims have far reaching implications. It involves the disruption of at least two major industries: traditional autos and oil. Both will peak and fall into decline within twenty years. It is critical that Tesla asserts itself to provoke this disruption. Without such provocation the peaks of these industries could be forestalled by decades. However, competitive situation is asymmetrical. The auto industry cannot compete with Tesla without hastening the end of conventional autos. Competing with Tesla cannot harm Tesla over the next ten or more years; it can only prevent Tesla from acquiring more than about 25% market share in the post-ICE auto market.
So let me set up this model. First, I segment the auto industry into four types of vehicles: Conventional (ICE), Hybrids (ICE+Batteries), EV (all electric, excluding Tesla), and Tesla. I will make the assumption that over time there will be a net market share transfer from C to H, E and T, and from H to E and T. It is important also to assume no net transfer between E and T. This allows all market share currently enjoyed by C and H to flow ultimately to either T or E. Both T and E will be absorbing states only if one assumes long-run parity , that there is no net transfer between them. I am also assuming that there can be no long-run parity between H and E. The rationale here is that battery costs will eventually decline to such a point that it is cheaper to add extra battery capacity than to add a small ICE for extended range, and other issues may obsolete all ICE even sooner.
In 2013, 82.84 million autos were sold world-wide. About 2% were Hybrids. Tesla sold 22k, and Nissan sold 90k globally. I assume that there were perhaps another 20k competitive EVs sold. So, I assign 110k initially to the E segment. This puts the industry at a 5 to 1 advantage over Tesla within the parity EV space. If both E and T were to grow at identical rates each year, this ratio of market share would hold, leaving Tesla with 1/6 of the post-ICE market. So basically the auto industry would have to be just as committed to growing EVs to retain 5/6 of the market. Even without a mathematical model, this is a pretty astounding observation. It is very hard to catch up with a competitor that is committed to doubling its business every year or two.
Disruptive Market Share Model
Let's make some formal assumptions about growth rates. First, we need some mathematical notation. Let C(t) be the number of Conventional vehicles sold in year t, and dC(t) be the change in vehicles sold from t to t+1. Apply the same notation to H, E, and T. Moreover, define the sum as N(t) = C(t) + H(t) + E(t) + T(t). Also define market share mC(t) = C(t)/N(t), similarly for the other segments.
Second, we assume that the market as a whole will grow at a small exponential rate, g, so that,
dN(t) = g*N(t), or simply, dN/N = g
I assume the rate g=2.75%. This rate is consistent with 100 million cars in 2020 starting with 82.8 million in 2013.
Next, we'll make two assumptions about segment growth that each share in general growth g in proportion to their size and that the transfer from one segment to another is in proportion to both segments. So, for example the transfer from C to H is the quantity
b_{CH}*C(t)*H(t)/N(t) = b_{CH}*C(t)*mH(t) = b_{CH}*mC(t)*H(t)
I point this out to emphasize that the quantity transferred is proportional to market share of either segments. This is a net transfer so b_{CH} = - b_{HC}. I call the coefficient b the transfer intensity.
With the above assumptions we can now derive the following growth formulas:
dC/C = g - b_{CH}*mH - b_{CE}*mE - b_{CT}*mT
dH/H = g + b_{CH}*mC - b_{HE}*mE - b_{HT}*mT
dE/E = g + b_{CE}*mC + b_{HE}*mH
dT/T = g + b_{CT}*mC + b_{HT}*mH
Note in the first formula that the growth of conventional is based on the general growth of the market minus growth that is being transferred to H, E, T. So long as the market share of these alternatives are small, then C can grow nearly at rate g. So it is that initially the incumbent technology can simply ignore challengers and focus on the bulk of the market. For example, if the intensity of transfer from C to H is 0.10 and the market
share of H is 2%, then C is only losing 0.20% in growth rate to H. That may be of concern, but EVs and Teslas are ignorable due to extremely low market share.
In the second formula, H is gaining general growth plus transfer from C, but losing share to E and T. Initially, mC is about 98%, so this growth rate is roughly g + b_{CH}. I set b_{CH} = 0.10 so that H grows at a little over 12% initially. This seems to be a comfortable rate for the industry. It's a modest amount of disruption. If the Hybrid market were to heat up and grow at 40% a year, it would grow to about 8% market share in 4 years and deprive Conventional cars about 3.20% in growth rate. This would overwhelm the general rate 2.75%, and the net result would be that C would have a negative growth rate. This would be disruption within 4 years. So long as the transfer intensity is low, it will be a long time until H disrupts C.
Finally, we turn to the two last formulas which are structurally equivalent. These segments are simply receiving market share from the other segments in addition to the general growth rate. Their growth rates differ by how competitive they are in capturing market share fro C and H. Since mH is so small initially, their growth rate is dominated by the transfer intensity from C. Tesla needs this intensity to be about 0.55 to grow from 22k in 2013 to about 500k in 2020. So this is my base case for this intensity parameter. It is difficult to know how much transfer intensity there may be from H to T. It's not exactly observable since the current market is dominated by C. It is reasonable to assume that it is a fraction of the intensity from C, since EVs and Teslas have bigger advantages over conventional cars than hybrids. So I'll simply assume half the intensity. One interpretation of this is that if the industry could switch all C into H, then the growth rates of E and T would be cut in half. It is possible that the industry may at some point respond to the Tesla threat by severely cannibalizing C with H. Even now, most of the new models touted as competitors to Tesla are in fact hybrids of one sort or another. While this strategy has the theoretical potential to slow the transition to EVs, the industry cannot or will not disrupt conventional cars fast enough. The growth rate of hybrids would need to be as fast as Tesla's growth rate. Arguably, that that effort would be better directed at growing EVs. The reason is that due to parity transfer of C to E would deprive T more growth potential than transfer to H. In any case, the disruption of C is inevitable the only question is how long it will take. So the key question is how much transfer intensity to E the industry will pursue.
Discussion
Tesla seems committed to an intensity of 55%, maybe as much as 65% or as little as 45%. Meanwhile, it is hard to imagine the industry mustering an EV intensity greater than Tesla. In my sensitivity analysis, we consider industry competitive intensities from 20% to 70%. Intensity of 20% would be a disaster for the industry, as T would walk away with around 95% ultimate market share. So while there may be such a low response initially, eventually the existential threat will motivate a higher response. That threat may become apparent when conventional auto sales begin to stall and decline irrevocably. The time of peak C will likely come by 2025 according to this model, but may come as early as 2022. Ironically, the higher the competitive intensity is, the sooner will come the peak. There may be some comfort in collective inaction, but there will also be really opportunities for those competitors that take up the challenge. The companies that establish strong market share within the EV space stand to capture a lasting hold on market share.
The disruptive market share model generates scenarios of how this could play out. I encourage readers to play with the worksheet and get a feel for the dynamics. I find it particularly challenging to come up with a realistic scenario that is favorable for traditional automakers. Conversely for Tesla, the strategic path is pretty clear. Tesla needs to grow by 45% or more per year, and it can walk away with 25% of the market. Actually, the longer it takes for the market to disrupt, the bigger its share. If the industry does turn up the intensity, then Tesla will likely be able to turn up its intensity as well. So Tesla is in the enviable position. It remains to be seen who may join them. A co-disruptor will need to come from outside of the industry or be willing to divest itself of its ICE business.
So have fun with the worksheet model, and let me know what you think. In future posts, I'll discuss model limitations and potential extensions.
View attachment 61227