SQNM Earnings Model Update & Projections
Finally, after a number of hours of digging through SQNM’s earnings release and reading through the earnings call and the EDGAR filing, I have updated the Excel file containing the earnings model. Here is the file. You will want to have this open as your read through the remainder of the article.
Q1 2013 Review
First, let’s review Q1 actual vs. the modeling predictions. The table below shows the estimation, what the actual number was in Q1 and the difference. The gist of it is that I was spot-on for the earnings-per-share (EPS) number – I had underestimated the total revenue and the total expenses by approximately the same about (roughly ~$5M), offsetting in the final EPS estimation. Not bad (if I say so myself) for a hobby analyst with no formal finance education!
|# of Tests||33990||44500||10510|
There are a number of assumptions in the model that were off, some more than others, and they will need to be changed to improve the accuracy of predicting the numbers for Q2 2013.
Assumptions: Past, Present and Future
# of Tests. Assumed: 3% Q/Q growth in tests accessed. Actual: 34.5% more tests accessed. The reason this didn’t throw my model of significantly (which it should have) is because payment was received for only ~35% of the tests accessed. The revenue recognition is on a cash basis, which means that even though the test was accessed/performed in Q1, Sequenom was not paid for it and, thus, did not include it in the revenue number. On the earnings call, management mentioned that it would change the way it recognized revenue to when Sequenom performs the test (accrual basis accounting), but that is not slated to happen until Q4 2013. Forward assumption: 10% Q/Q growth in number of tests accessed. Even though Q1 2013 posted much higher Q/Q increase, I want to play it conservatively and choose a less aggressive growth number in case Q1 was an anomaly.
Products Revenue. Assumed: 10% Q/Q revenue decrease. Actual: 24.6% decline in revenue. I underestimated the decline in revenue for this segment by nearly $2M. It appears that cannibalization by SCMM is proceeding more rapidly than I had anticipated. Forward assumption: 20% Q/Q decrease in Products revenue. This is an aggressive decrease assumption. My desire is to model a conservative scenario for Sequenom’s performance, a scenario below which the company should not perform if its business is headed in the correct direction.
Expenses. Assumed: flat Q/Q. Actual: 3.9% increase in cost of revenues and 3.5% increase in total operational expenses. While the increase in expenses was relatively small compared to previous quarters, it was still more than I had estimated. Forward assumption: 2% Q/Q increase in cost of revenues and 2% Q/Q increase in total operational expenses. Because Sequenom is growing its top-line at a frenetic pace, that growth will come with a slight and continued increase in operational expenses.
Revenue/Test. Assumed: 2% Q/Q growth in average sales price (ASP) per test. Actual: 2.2% Q/Q growth. The assumption was quite close to reality. As mentioned above, the ASP includes cash received for tests performed in the quarter and cash received for tests performed in prior quarters. We don’t have a more granular break-down, but as the year progresses, the company should see an increase in ASP, as that would parallel the increase in tests accessed since launching SCMM. Forward assumption: 2.5% Q/Q growth in ASP.
Other Adjustments. By by calculation, stock-based and other expenses shaved $0.03 off of the EPS. In Q1, the outstanding share count increased by 250k. While these additional factors have an obvious and significant impact on the bottom line, they are notoriously hard for me to predict. Going forward, I included them in my model at a flat $1.5M per quarter in stock-based and other expenses and increased outstanding share count by 100k each quarter.
Q2 2013 Predictions
The application of the forward assumptions to the Q1 2013 numbers leads the model to yield an EPS estimate of -$0.24 for Q2 2013. This number is $0.02 below the consensus of nine analysts currently following Sequenom.