Benchmarking Multi Product Adoption in Software: $DDOG dominance, Databricks making history, Apps vs Infra 🧵
💭 New product adoption is critical for revenue + margin expansion + retention. This latest analysis benchmarks new product penetration of ARR vs time since launch & can be a helpful framework when trying to underwrite new product penetration in Software.
*New data set includes disclosures of 17 public software vendors, including new data from $DDOG, $CRWD, $WDAY, $ZM.*
🔑Key Takeaways:
1) Average annual new product penetration of ARR across the 17 vendors is ~3% per year

2) Infra Software new product penetration of ARR is close to 2x that of Apps Software companies. Infra companies tend to be more PLG and bottoms up driven which could explain the difference, in addition to the vitality that products tend to have with developers, vs front office app end users.
3) $DDOG clearly stands out and their disclosures speaks to their ability to expand into other areas of observability. The company was able to get APM + Logs (although 2 separate products) to ~45% of ARR in ~5 years post launch. Given the disclosures from the company are including 2 products, it’s fair to assume they are similar/equal in size, meaning $DDOG has gotten each product to ~22% of revs in 5 years (~4-5% per year).
4) Databricks new disclosures show that their Datawarehouse product is approaching ~17% of ARR, only 2.5 years post launch of the product. *Based on all public data, this seems to be the fastest new product launch in terms of ARR penetration in the Software landscape currently*
5) $CRM was able to take Service Cloud to 25% of revenue 6 years after the launch in 2009. This makes Salesforce the most successful/fastest penetration among Application Software companies (~4.5% penetration of ARR per year). Service Cloud was a natural adjacency for Salesforce to expand from Sales Cloud.
6) $WDAY launched ERP in 2008 & it took ~8 years for the company to get to ~8% of revenue. $WDAY’s expansion into ERP was very hard as the buying center for HCM was completely different than the one for ERP. The acquisitions of Adaptive/ScoutRFP added ~7pts to ERP penetration of ARR in 2018/2020 respectively. Today ERP penetration of ARR has stalled out at ~25% (similar to CRM’s penetration ceiling of ServiceCloud), an average annual penetration of ~2% per year.
7) $ZM was able to leverage its massive Video installed base by getting Zoom Phone to ~11% of Revenue in ~4.5 years. $ZM chose to massively discount this product vs the market, with an implied $6 per seat per month vs the industry’s $15-$20 per seat per month pricing.
🔗 Full Link to Analysis Below
Going through expert network transcripts & pretty glaring how poorly buyside uses expert networks
The amount of analysts asking “experts�� what they think about next year Street growth & margins is shocking
These “experts” literally have no idea…that’s your job to figure out
Coming around to the view that the lower the amount of expert calls I’ve done on a name, the more alpha I’ve made on that name
Sometimes going too deep creates confirmation bias
There just not much differentiation today by having that be a big part of process
Less is more
That being said I’m definitely listening to the sellside checks call with a partner that does $10k in sales for a vendor that does $1b in revenue 😂
Best way to get employee productivity up inside of a company is to cut WiFi access in the bathroom
Especially if 5G signal is very low
You’re welcome Tech CFOs
@MadsCapital Not much different from sellside that claims they do all this deep work when they just download excel from Gartner and IDC as their “TAM analysis” 😂
@bucketshopcap Biggest change and lesson has been to fade extremes in non large cap stuff - whenever too much optimism (this is 100% winner, best product) or too much negativity (this is a 0, AI loser).
Usually the answer is somewhere in between but market aggressively prices in one case
5 yrs ago everyone thought $ORCL database business would be declining, seeing meaningful churn to newer NOSQL database technologies
Fast forward to today, $ORCL disclosed at their Analyst day that the Database business has an net retention rate of 102% (includes px increase)
@when_to_fold_em Pod: “Days adjusted growth for next quarter is .2% below the 4 year stack. Why is that?”
Vs
LO: “Where do you see Stock based compensation going in 10 years”
Choose your fighter
It amazes me how bad 99% of sellside models are, especially since it’s an inherent product & cost buyside pays for
No color coding
Revenue forecasts that take last quarter growth & continue it forever or just hardcode numbers
Hardcoded numbers on cash flow
Gridlines shown
50 random tabs with nothing material in them
Why do you have 6 associates on payroll?
Why is it so hard to have an easy working model your clients can use?