This is not our work, so I take no credit for it.......nor any...

This is not our work, so I take no credit for it.......nor any criticism .

Volatility of ARR Growth at LiveTiles

Management have mentioned that quarterly ARR growth can get affected by the closing dates of large contracts. We examined quarterly ARR growth and growth in number of customers for signs of large customers impacting quarterly numbers. If a few large customer wins falling either side of a quarter end are significant, the first sign of this would be some volatility in the size of the average customer win which coincides with a quarter being above or below expectations. We would expect to see upside quarters coincide with a big increase in the size of average customer (because some very large customer closings fell in this quarter) and downside quarters coincide with a decrease in the size of average customers (because very large customer closings fell in the preceding or subsequent quarter and not in the current quarter).

Stripping out the customers added with the acquisition of Wizdom (disclosed in the LVT annual report), we observe that March and September 2019 had average customer win size that were significantly smaller than the significantly larger adjacent quarters suggesting that large closings had occurred in adjacent quarters.

Modelling a "best fit" to explain quarterly ARR growth, we believe that LVT has a large and relatively consistently growing number of median customer wins each quarter but also a small number of very large customer wins of $500,000+ which vary significantly from quarter to quarter. These very large customer wins have a significant effect on quarterly ARR numbers. The mathematical best fit is a median win of $70,000 over the past year accompanied by large wins of $600,000 . If these large wins were distributed evenly over the year, LVT's growth would more closely fit growth models. We are not concluding that LVT has $70,000 customers and $600,000 customers but the close fit of the model reinforces the observation that the business has had a mix of steady growth from mid size customers + lumpy growth from very large customers over the past 5 quarters.. We postulate that merely 2 fewer large customer closings that did not occur Sept 2019 vs Sept 2018 may obscure steady y on y ARR growth in median customers.



It is interesting. Note the model was not forecasting the ARR, it is simply a best fit model to explain the historic ARR. You can see the similarity to Mags work and perhaps this is useful for Mags model.

Personally I think any model based on # of customer wins and customer size going forward will have to be significantly tweaked because :
1. The company is embarking on a renewed effort on a partner channel. So one would have to model how many customers will come through a partner channel and what the average size of those customers will be net of partner revenue share.
2. The Direct sales force is focusing more on the so-called "Mega" customers . One has to model whether this will increase from 12/14 per year to something (pick a number.....presumably the company has a target).
If I saw a model that took account of this, I would be interested.