Playing with Numbers: What a Microblogging Service Financial Model Might Look Like

Building a business out of a microblogging service is not a trivial task. There has been a lot of chatter regarding Twitter’s eventual monetization, but so far the only people who made money off microblogging are the Jaiku founders when they sold the business to Google. I have written before about the difficulty of monetizing microblogging, but difficulty is something to overcome, not a brick wall. This model is an attempt to figure out the different options of running such a business.

You can try the Google Docs version (if you want to play with it make a copy first). If you know your way around financial models, jump right into the model. Continue reading for some background and points of interest.

Financial Model Snap

When Nouncer was planned to be a consumer-facing microblogging service, just like the rest of the pack, I needed to get an idea of the numbers behind it. The most important part was trying to figure out the cost, mostly from a funding perspective. I could not find any good model out there to use, as microblogging has a lot more back-end processing than front-end (that whole messaging system). I also needed to learn about SMS and its cost structure. The way I approached it, was to make a list of all the costs and revenues components of the business, and for each item, ask myself what properties can I assign to give more depth to the bottom line.


  • Infrastructure – servers, bandwidth, storage, facilities. What patterns define how the service is used? What is the nature of the messages, their size, and frequency? When trying to forecast SMS costs, what percentage of the users will want to receive messages on their phones? How many will update their information from their phones? What does a minimum deployment looks like? How many users can a server handle? What’s the ratio between registered users to non-members visiting the site?
  • Staff – salaries, recruiting. What are current market prices for different roles? What is the pace at which new employees are needed? What are the costs associated with employing staff?
  • Business – office space, legal / accounting services.
  • Banking – transaction fees (the cost of receiving money).
  • Advertising – user acquisition cost.


  • Membership – tiered accounts. How will the user-base grow? What percent of visitors will sign-up as members, how many will buy a membership and at what level? How much more active are they likely to be? What is the expected user renewal rate each year? What to charge for membership? What features to offer in each level?
  • Advertisements – web, SMS, instant messaging. What to charge for each type of ad? What click-through rate to use?
  • Directory – listing fees (for promoting content). Type of listings? Cost? How many members will buy them?

The spreadsheet structure contains a few areas: the top is the control panel with all the variables grouped into categories (green for revenue-related, red for cost-related, and blue for user patterns), with the 5 year projection below, followed by the fully detailed 5 years month-by-month breakdown. All the math is done in another sheet. In the control panel, blue numbers are changeable and will impact the bottom line.

Some interesting points:

  • Membership Growth – defines the growth graph by giving a different growth rate for different time periods. For example, the model expect 100% monthly growth in the first 3 months, down to 50% in the following 3 months. The model expect growth spending after a year at $0.24 a user for 6 months, then $0.36 a user. The numbers are based on being able to start the model with 5000 members (it is hard to model growth in the first few week and getting an initial 5000 users isn’t too hard).
  • Revenue Conditions – minimum number of users (critical mass) needed to start charging for services and selling ads. Also lists the number of months of development before the system is ready for the first user (5 months).
  • SMS – has the option of reducing SMS cost by using free email gateways (try is – see what happens to the bottom line when you don’t use email for SMS). While this is not 100% kosher, it is still a viable option which may or may not be allowed by the phone companies. Either way you need to be successful for them to notice you. SMS costs are based on data from 2006 and has a lot more explanation in the comments.
  • Staff Cost – sets the minimum number of people needed for each role (sys admin, support, developer, and manager). Managers are assumed to be founders and the model does not account for their pay. The panel provides control over the staff growth with the radio of each role against the number of users it can support. For example, one support staff is needed from the beginning, they get paid $60K annually (actually less as these numbers are employer cost), and each support person can cover 300K users until the service reaches 600K users – at which point each support person can handle 400K users. The two tiers are needed to better control staff growth in the early stages when people don’t scale that well.

I am no longer using this model as the nature of the Nouncer business has changed to a B2B model providing services for others to build consumer products. Feel free to post questions in the comments or any other feedback.

One thought on “Playing with Numbers: What a Microblogging Service Financial Model Might Look Like

  1. Fabulous. Thanks, will be downloading this now.
    While PassPack isn’t a micro-blogging platform, and we essentially have our model worked out, I’m very interested to see the details of how some of your logic and assumptions compare to ours.
    Thanks for being open and transparent. I hope others will benefit.
    Cheers – Tara

Comments are closed.