Grocery delivery in 10 mins?
The past few months have seen a deluge of startups offering grocery delivery in 10–15 mins, especially in Europe. London alone boasts of more than 7 of these “ultrafast” grocery delivery players, the most notable (and heavily funded) being Weezy, Dija, Getir and Gorillas. In Q1 of 2021 alone, they’ve raised over a Half Billion Dollars in funding.
So, why the massive excitement for this service?
Well, the world is moving towards “instant gratification”. If you don’t need to wait for fund-transfers, getting social validation, talking face-to-face with your friends or colleagues, why plan ahead and wait for groceries?
With this insight, grocery delivery has now moved beyond beyond disrupting “supermarkets” (monthly/weekly planned purchases) and has entered the realm of “neighbourhood convenience stores” (unplanned/ need-it-instantly purchases). These 10-min players, are then effectively becoming the seven-eleven of the online world.
There are strong headwinds for such a service to gain adoption. As remote working becomes common (even post-pandemic), opportunities to “pick up a few things on your way back” reduce, and the allure of these 10-min deliveries grows. In addition, new technologies like AR/VR on top of digital communication (zoom/slack) will expedite this “nesting habit” where people would increasingly step out of their homes less often. It is important to note that COVID, while the major catalyst for online grocery delivery, is only accelerating an already existing trend — which is here to stay even post pandemic. Instant doorstep delivery is just one part of the charm of these services, the other part is potential access to a wider assortment and high predictability over inventory (you don’t need to physically go to the nearby seven-eleven only to find out they don’t have your favourite soda available).
In the future, as people get hooked onto this “comfort” of not having to plan grocery shopping in advance, these 10-min players might start capturing more of the monthly/weekly purchase use-cases as well!
Is 10 min delivery even possible?
Yes, it is — but it isn’t easy. As 10-mins hardly leaves any room for error/ delays, these players control the process end-to-end with their own dark stores (almost all players), and their own delivery fleet (instead of gig-economy riders on UberEats, Deliveroo).
In order to deliver majority orders within 10 mins, players need to ensure
(1) Extremely fast order-picking (less than 2 mins). Dark stores employ several innovative picking models here (order/batch/wave/aisle picking)
(2) Restricting delivery radius to ensure rider needs not more than 8 mins to deliver order. This also means players need to set up a high-density of dark stores (40+ stores in London) to cater to majority of the population.
(3) Ensuring near-zero time from “order-received to start-picking”, “order-packed to picked-by-rider”. This involved building buffer picking capacity in the dark stores, and rider capacity so that there is always a picker and rider ready for any order.
Enabling this high-bar on customer experience of course makes this a more expensive operation than same-day/next-day grocery delivery through larger city-level fulfillment centers. Further, because of the need to manage several dark stores in the city instead of single central fulfillment-center, supply chain economics of this model are poorer compared to the same-day/next-day delivery models. This leads us to the next question — is this even economically feasible?
Can 10-min delivery make money?
The answer, according to me, is Yes again. However, it is a difficult tightrope walk — players will need to perfect planning and execution on several levers (inventory planning, supply chain operations, dark store operations, wastage management, rider demand forecasting, pricing).
Below is a potential order-level unit economics snapshot:
A few imperatives emerge from this snapshot above:
- There will be a premium associated with the convenience of 10-min delivery. It will manifest itself in item prices and margins (hardly any discounts on suggested price). Thus the target audience for this will not be value-seekers who are willing to buy in advance/in-bulk/on previous day to get discounts, but rather convenience-seekers who are willing to pay a little extra.
- Most players will also charge a service fee (to add to revenues and to ensure higher order-values/ dissuade customers from placing single-item orders). However, in a healthy unit economics, the revenue contribution of this will be lower compared to commissions and other sources.
- Other sources is an important revenue line-item for these players. These will largely be FMCG brand ads/ sampling/ other monetization avenues.
- At 2–3 GBP per order, rider cost will be the largest cost-element — maintaining high utilization levels for riders will become critical, else this figure can really shoot up.
- Supply chain costs (especially wastage expenditures) can also really explode and wipe-out the entire unit economics. Another related cost element is refunds/ penalties for wrong orders. Both these point at the need for very sharp operations management.
One important cost element missing from this analysis, is “fixed costs” (mostly rent and maintenance costs) for the dark stores. Given the need for high density of dark stores, these costs could go prohibitively high and make the model infeasible.
What about the investment in setting up
Typically, ideal geographies for this service would be cities with high traffic congestion (making deliveries a great alternative) and high population density (potentially ensuring a sufficient customer-base within a 10-min delivery radius).
In fact, the latter condition could potentially limit the number of cities where this service is economically feasible. Given the high fixed costs of these dark stores, there should be sufficient order volume at a reasonable unit-economics for every order for every dark store. I applied these conditions to identify what population density would make this feasible.
As per these calculations, this model is feasible only in locations with a population density of 4000 people/sq km. Now I used data for London for this calculations, so as you go to other cities, the following things change:
- Rental costs/ Sq Ft (I assume them to be half of London in other cities)
- Traffic and hence, the delivery radius which I can service within 10 mins
- Labour costs and propensity to pay delivery charges (I assume these to neutralize each other, and have assume the same unit economics across cities).
With these changes as well,
This model is feasible only in cities with a minimum population density of 2000 people/ Sq Ft.
Applying this filter on all cities in Europe, only 32 cities (below) qualify. This explains why cities like London are prime targets for so many of these startups! The total grocery market of these cities is estimated to be over $300Bn. So all these players are battling for just a slice out of this pie!
Of course, these are back-of-the-envelope calculations, and the number of cities could be lower (say, 20) or higher (60–80). But the important point here is, it will likely be a very limited set. In fact, Getir the Turkish pioneer in this model decided to expand to markets like London, Sao Paulo, instead of going beyond the top 7 cities in Turkey.
Another insight from this rough calculation is that because each player needs to have a high order density,
This model can only support 1–2 players in any city eventually.
So 10–15 min deliveries should eventually see some consolidation as competitors see they haven’t gotten sufficient traction in those geographies.
Who will ultimately win this war?
While it is too early to pick a winner, there might be some clues into what would be a “winning strategy”. Any growth company needs to focus on three variables (1) Growth, (2) Unit Economics, and (3) Customer experience.
As the underlying products that they’re offering are not really differentiated, I strongly believe that the winning strategy would be double-down aggressively on Growth (1) to acquire customers before the other players do, and have an even stronger focus on superior customer experience (3) to lock in the customer (evident in retention figures). Of course, there would be guardrails on economics, but positive unit economics are unlikely to play out till a player reaches significant scale (close to full-penetration of market, 50%+ share) and not something to solve for at the beginning.
A player who is set up well for growth would have
- Sharp understanding of the assortment strategy (this model can support only 1500–2000 SKUs)
- Early and rapid investments in brand and distributor partnerships to ensure high availability of stock, and rapid additions to assortment as new products get introduced
- Demonstrated ability to expand rapidly into new neighborhoods (setting up dark store, and a large delivery fleet)
In addition, companies should have figured out the most effective growth loops for themselves, which translate to predictable and reasonable customer acquisition costs.
Ensuring high quality CX would require strong operating processes to ensure
- Adherence to 10-min delivery time: timely picking, and delivery
- High fill-rates: close to 99%
- High quality: one incidence of rotten veggies/ broken/ expired packets is sufficient for the user to switch to another service
AND, investment in a highly trained and responsive contact-center for customer complaints, with a sharp focus on customer-retention in the face of orders gone wrong.
Lastly, one important factor to note, is that every new city is a new battleground — success in one city does not give a distinct advantage in entering a new geography. Barring a few advantages (FMCG partnerships, Brand, app and digital assets), companies will have to set up the entire infrastructure from scratch. Hence, a high-potential player would have built a strong playbook and team to replicate their success in more cities.
A look into the future
I am very excited to see how expansion plans for these companies play out over the next 12 months. It is very likely that a year hence, we see several players who have gotten phenomenal traction in a few cities, but haven’t managed to build a stronghold in others (e.g. Getir in Istanbul, Weezy in London, Gorillas in Berlin). Being the king of a single city might not justify the millions of dollars of investments in these companies — and we should expect some consolidation, even among players who have managed to capture a few marquee cities.
On another note, the extremely low picking times, and need for very high accuracy point towards a solid business case for automation in these dark stores. While most current microfulfillment automation solutions are not suitable yet, for these dark stores, it is only a matter of time — and I am truly excited about opportunities in that space!
