Crowdsourced competitive intelligence: InfoArmy

A new website launched today,, touting itself as crowdsourced competitive intelligence. At launch, 94 company reports are available, and clearly the goal is to have many hundreds.

infoarmy competitive intelligence.png

Each report is $99, which seems reasonable, and at the same time manages expectations as to what you can expect - these may be good summaries, but you can only expect limited research at that price. I cannot say what these reports look like since, strangely, reports are only viewable on the iPad (InfoArmy tell me that PC downloads will come later). I know that enterprise IT is being 'consumerized' but surely most business people will need these reports on their work PCs/laptops.

Anyone can write a report. These are reviewed by Senior Researchers, who are researchers with seven published reports. Presumably these are in turn reviewed by InfoArmy staff, especially in the early days. The main researcher gets 40% of revenues, the senior researcher gets 10% and InfoArmy gets 50%. So if you write a report a month and each one sells once a month, that's an annual income for the writer of $475.

The InfoArmy forums are a good window into how the reports will be researched, with interesting discussions about techniques and ideas. For example:

  • When estimating the revenue for a company, one of the easiest ways to develop an estimate is to start with the number of customers. Next, you can either locate information on revenue per customer per year, or infer based on the company's product offerings. I created a Google Doc that describes this method and demonstrates an example: here.
  • Besides the obvious (LinkedIn), Angel List is a good site for execs and board members, especially in the tech and VC world: Also, don't forget to use LinkedIn, etc. to find executives who have LEFT the company. This is important for the People Timeline in the report.
  • [to estimate revenue split by product] You need to look for clues through the intelligence that you are gathering along the way for your respective profile. As it's unlikely you'll find a solid breakdown for each company you review, you need to try to piece snippets of data together, with a view to supporting your estimates. Obvious starting points include the launch dates for products/when they came out of beta (company blogs/press releases/industry reviews etc). Are products being positively reviewed alongside peer products? Are they next generation products? If you're considering an app of some sort, you need to check whether it's being offered for free, whether it's cross platform, or for example restricted to the iPad (to consider how wide its audience reach is). If you happen to be reviewing a portfolio comprising 4 products, it is very unlikely that each of these will be contributing a neat 25% to revenues.
  • Cubic spline interpolation is a good way to fill in "the gaps" in your revenue data if you have at least three good yearly revenue figures, two of which are endpoints (earliest and last years). This method will not help you to extrapolate data beyond your last year of data or prior to your first. Other modeling methods could be employed to do this, one of which - exponential growth modeling - I'll cover in a future post. The Excel spreadsheet you may download from this link is an example for Amazing Charts, Inc. that I worked up. Amazing Charts licenses one of the best electronic health record software systems for small medical practices. There are extensive notes in the sheet (go down below), as well as explanatory comments in various cells. The main purpose of using the cubic spline approach is to derive estimates for the last 6 quarterly revenues. A key note here is that the actual data to "spline" is the total cumulative revenue from year one, not the yearly revenue itself.

For new researchers (and even old hands), these discussions may develop into a very useful resource about the estimates and guesswork involved in building competitor profiles. The quality of such advice, and perhaps more formal training, will be important given that anyone can be a researcher.

It is early days, particularly until these reports are downloadable to PCs, but this may be an interesting service to watch.