BlogBlog JungleworksBulbul

10-Step Framework For Evaluating On-Demand Startups

By Mahesh Vellanki 22nd December 2015

Framework For Evaluating On-Demand Startups

This post was first published on Mahesh-VC.com

There are many exciting on-demand start-ups out there going after some really interesting market opportunities.  Sadly, many of them won’t make it in the long run. In the previous generation of marketplaces there were countless contenders going after many different opportunities.  Yet, only a few became highly successful – companies such as Yelp, Zillow, GrubHub and TripAdvisor.  My guess is that the same will play out in on-demand.  In order to evaluate the ones that have the potential to succeed, my colleagues Ryan Sarver, Archit Bhise and I put together a 10-step framework that guides how we evaluate on-demand companies. We presented this framework a round table event recently hosted by The Information.  The ten categories are:

  1. Vision
  2. Market Size
  3. Go-to-Market
  4. Unit Economics
  5. Retention and LTV
  6. Supply Side
  7. Operational Excellence
  8. Market Growth
  9. Network Effects
  10. Technology Leverage


Is there a cohesive vision behind the company that goes beyond “Uber for X”? What is particularly transformative about the service? Why now? Why hasn’t this been done already? What could the service ultimately become and how does that change the market opportunity? These are the initial questions that need to be understood before digging into anything else.

Market Size

A lot of companies quote huge market sizes as if that makes the company inherently exciting. We find the exact opposite to be true. Going after a “massive” $350B home services market may be far less appealing than going after a “broken” $10B taxi market in the case of Uber. When going broad, you’re literally competing against every vendor that touches that market, both offline and online. However, by tackling a particular problem in a more narrowly defined market, a start-up can become the new standard in that market which can create significant moats or consumer behavior change. It is the classic red ocean vs blue ocean discussion.  The blue ocean businesses have a better chance to exhibit monopolistic characteristics which make for better business models.  In short, don’t fall for what we call the “broad TAM trap”

online and offline provider


Paid customer acquisition is a common strategy for many consumer start-ups. While many large companies leverage their marketing muscle to create moats around their business (eHarmony, Match, Geico), paid customer acquisition is not a strong differentiator for start-ups. That is, start-ups typically use well known channels and acquire users for comparable amounts. A better strategy is to evaluate whether there are any creative marketing strategies, viral hooks or other organic growth strategies that can advance your position. What is it about the service that lends itself to strong word of mouth or brand affinity? It also helps that organic users generally perform better than paid users.

Unit Economics

Even in strong on-demand models, you need to drive much higher top-line growth than in traditional two-sided marketplaces in order to drive a similar exit return. The inherently less profitable margin structure of on-demand companies underscores this point. Marketplaces such as GrubHub and Zillow can generate upwards of 30-40% EBITDA margins at scale, while on-demand equivalents might generate somewhere in the range of 5-10% (if that). Thus, on-demand businesses need to drive their top-lines much higher. Can the on-demand business take enough of a cut from the supply side or extract enough from consumers to make the business viable? Are all the costs being properly accounted for? Investors and founders should be particularly weary of burying variable COGS “below the line” into operational costs since this paints a rosier picture than reality. Below is an example of a transaction for a GrubHub-like company compared to two on-demand models.

on-demand food delivery comparison

Retention and LTV

Unlike a good SaaS business, consumer businesses tend to have fairly extreme retention curves. It’s not unusual to see tons of customers flow into the service and then churn out quickly. This is not necessarily bad assuming that the value of the remaining loyal customers make up for the churned ones.

Retention curves of saas platform
Retention is the first half of the equation. The second half is frequency and order value (taking into account margin).  Put these factors all together and you can begin to understand LTV and payback. All of these things vary drastically depending on the type of service you are providing. I’ve provided some illustrative examples below.


Supply Side

Like most marketplace businesses, on-demand start-ups are generally supply constrained. Having tight supply side operations is crucial for on-demand start-ups especially given the competition for contract workers.  Initial costs for recruiting and on-boarding labor should be properly accounted for. Some portion of ongoing recruiting costs should be allocated as cost of goods sold, not buried “below the line” as a fixed cost. We expect high churn but hopefully not 100% churn.

challenge supply side retention

Operational Excellence

Operational rigor is critical given the tight margins related to managing a networked labor supply. Operational hiccups can quickly cause chain reactions: unhappy labor –> high supply churn –> lower service quality –> unhappy customers –> higher customer service costs. How is technology being used to streamline operations and gain economies of scale over time?

Market Growth

It’s great if the home market is growing nicely but what about new markets? Are lessons learned from home markets being applied to new markets successfully?  What are the pros and cons of each market?  How does the service adapt? What can a single market be worth at scale? It is vital to track the progression of each market and the company should be really hands-on with individual P&Ls by market, market launch roadmaps, and benchmarking key metrics.

revenue in month since launch

Network Effects

An incremental new user should improve the service for all others users and each new supply-side addition should create better liquidity. What other key moats and competitive advantages exist? How is the service differentiating itself – is the value prop focused on cheapest price, highest quality service, or perhaps best liquidity or selection? The topic of network effects has been written about at length so I won’t belabor the point here.

Technology Leverage

For on-demand start-ups to get valued like technology companies they need to, well…demonstrate efficiencies from technology! Otherwise, they may just end up looking like massive offline agencies. Every part of the business should ideally be automated with technology. At scale, technology should help to improve unit economics and reduce costs and other friction points in the marketplace.

Parting Thoughts

We’ve learned many lessons in on-demand over the past few years. If we were to provide some cautionary advice to new on-demand start-ups I would shrink the learnings to just a few bullet points:

  • Frequency matters in mobile and helps build habitual behavior.
  • Be careful of hidden unit costs and poor accounting.
  • Look for technology leverage and economies of scale in the model.
  • Be weary of simply becoming a more efficient agency.
  • Be very careful of discount customers.  Can prop up short term growth, but not sustainable and can cause negative chain reactions that are hard to fix.
  • Look for teams that are equal parts star operators and technologists.

We’re really excited about on-demand companies that are creating magical experiences for consumers. The ones that are able to successfully navigate the 10-step framework above will have the best chance at reaping the rewards.

This post has been written by Mahesh Vellanki who is a growth stage investor with Redpoint Ventures. 

Subscribe to stay ahead with the latest updates and entrepreneurial insights!

  • Share this article:

  • Blog Jungleworks Blog Jungleworks Blog Jungleworks