The rockets are taking off…

Accounting Platforms as Moats

Companies and people are excited about AI, Machine Learning and Deep Learning. What impact does this all have on accounting software?

Late in 2015 Andrew Ng the Chief Scientist at Baidu did a presentation on Deep Learning. His 20 minute presentation is technical but early on he compares Deep Learning to rockets. He explains that getting a rocket into space requires two core ingredients, powerful rockets and lots of fuel.

In Deep Learning, the powerful rockets are computers which in 2016 are ubiquitous (via services like AWS). Fuel is the second ingredient and that is likened to data. Lots and lots of data.

Engine and fuel. Computers and data.
Engine and fuel. Computers and data.

So why liken Accounting Platforms to a moat? Well over time the larger an accounting platform becomes the more challenging it is for new entrants to compete with brand recognition, complex feature builds, education of the market and more. On top of all that the vital attribute going forward is data — lots and lots of high quality accounting data in a central location.

Most people are guarded with their financial information and the result is very few central stores of financials. In fact there are no other platforms with access to the data online accounting software companies have. Let’s look at the alternatives.

Banks have data but it is limited to your bank transactions. No details on GST/VAT, no breakdown of payroll data, no understanding of each transaction, no understanding of inventory. Similarly a government tax agency has a lot of data but it is mostly a value under a label like “all sales made this month”. Even an eCommerce platform has sale data, possibly some purchase data but it’s still missing a lot of data like bank reconciliation data, payroll, expenses and more.

Contrast this to accounting systems that have or have access to:

  • every detail of every sale
  • every detail of every purchase, expense
  • all payroll details
  • all inventory details
  • every single bank transaction, individually categorised
  • details of all your customers, suppliers
  • details of all your legal tax filings
  • and more…

Accounting platforms have this data for 500,000+ companies and growing. Why is the 500,000+ number important? Andrew Ng’s presentation shows only one slide twice and that is this one:

Scale.
Scale.

AI & Machine Learning advances are flowing through larger organisations like Apple, Google, Facebook, Baidu and more. This technology will soon drive new accounting software features at which point the first set of accounting rockets will take off. Just hope you get there quick enough as once the rockets take off they are hard to catch.

Note: for those interested here are Andrew Ng’s slides: http://www.slideshare.net/ExtractConf/andrew-ng-chief-scientist-at-baidu