Nstein, a Canadian text mining vendor, who — through a combination of acquisition (e.g., Eurocortex, Picdar) and licensing (TextML) — has re-invented itself as a provider of solutions for newspapers and magazines, announced its 2Q08 results yesterday.
(All figures CAD$)
- Revenue of $6.0M
- Slight decrease in sequential revenue from $6.03M in 1Q08. This is unusual from 1Q to 2Q for a software company.
- Year over year revenue growth of 50% compared to $4.1 in 2Q07. This looks good, but is half inorganic (see below).
- EBITDA of -$1.1M, up from -$0.5M in 1Q08
- Net loss of $1.6M, up from $0.9M in 1Q08
- Cash burn of $1.8M
- Ending cash of $7.7M
License revenues were lower than expected, resulting in increasing EBITDA and net losses:
“This setback is explained by a lower level of licensing revenues than the Company had anticipated, since some customers delayed their decision to purchase.”
Note that the “slipped deals” card is one that should be played carefully because, technically speaking, if deals really slipped and everything else is on track, then analysts should add the value of the slipped 2Q08 deals to their existing 3Q08 estimates. But I wouldn’t count on it.
On the surface, revenue growth looks good, until you see that half was inorganic :
“Equal credit for this growth can be attributed to the acquisition of Picdar in February 2008 and the organic increase in the Company’s revenues.”
That means 2Q08 organic growth was 25%, down from 59% in 1Q08 and 61% in 4Q07, assuming those figures are purely organic which, as far as I can tell in ten minutes, they are. (Once an acquisition has been on the books for 5 quarters it no longer provides any “purchased growth” effect because both the year-ago period and the current period include the acquiree’s revenue.)
My primary beef with Nstein is they appear to sell a lot of services for a company whose value proposition is all about pre-integration. After all, if you wanted an integrated XML server, CMS, and DAM system, you could integrate MarkLogic, Documentum, and Artesia and have a lovely best-of-breed strategy.
Nstein’s argument is that that’s too expensive to buy and integrate. So, instead, you should use their pre-integration of three products from companies that they have, in cases, acquired for less than a single large Documentum license deal (e.g., Eurocortex cost $1.4M).
Overall, I call this the “moldy sandwich” strategy. If customers want sandwiches:
- You can get crisp bread, fresh ham, and home-made mayonnaise, and ask your customers to do the assembly
- Or, you can get moldy bread, old ham, and sour mayonnaise — but pre-assemble the sandwiches
Putting component quality aside, vendors following the second strategy should require relatively few services to deploy their products because the sandwich is, after all, already built.
Now Nstein doesn’t break out services revenues, but by knowing typical software vendor gross margins by line of business (i.e., software, maintenance, consulting) and doing some reverse engineering, you can build a model to guestimate what things look like. The big clue is the 52% overall gross margin, quite low for a software company (norm 80% to 85%), which suggests there is a large amount of relatively low-margin consulting business. See below:
Note that this is just one way to make the figures work, but the simple problem is that software gross margins are so high that there’s simply not room for much software revenue in the overall mix, and it’s hard to drop maintenance (also typically high margin) below 20%. Even if I model consulting as losing money, it’s hard to get license above 40%.
What does this mean? If you’re selling pre-assembled sandwiches, then why are you selling $3.3M in services per quarter? Could it be they aren’t as pre-assembled as they’re pitching?
And if they aren’t, then what’s the value proposition again?