AMPL: A closer look

With several members on this board taking sizable 5/10% positions in Amplitude (AMPL), I started digging in the archives to form an opinion on this company and it seems no comprehensive post was presented for what looks like a very attractive opportunity

So, here’s a quick primer on AMPL

For context, the company directly listed (instead of IPO, will come back to this) on 28 September 2021 at a reference price of 35$ and closing the day at 54$.

What does it do? Amplitude is a product-optimisation SaaS tool that enables customers to get insights on how their digital products are used in order to make data-driven decisions on how to improve their consumers’ experience

Say it in English: With Amplitude, a company like Peloton can easily find out that the customers who are much more sticky are those that engage with a certain sharing feature currently hidden three-screens deep in a menu. They can hypothesise that by placing that functionality on the home-screen more customers will engage with and therefore return more often to the app. They can carry out an A/B test to verify this hypothesis and, once confirmed, they launch the change driving meaningful increase in revenue. This clear ROI is a cornerstone of the Amplitude offering, as it is the foundation for their usage-based model

Who do they serve? Historically they were focussed on SMBs, but after years of learning they now successfully penetrated the Enterprise segment, with 26 FTSE100 customers in Q3 2021 and 1417 customers in Q2 2021 of which ~25% drive $100k ARR or more. They also support a large and generous free tier that enables customers to trial the product until they are ready to upgrade their usage to a paid plan. This is a playbook we have seen Cloudflare deploy very successfully, both in terms of Go To Market but also to enrich their datasets and have insights into new product development and beta-testing new offerings
How big is it? The company is in it’s early innings, with Trailing Twelve Months revenue of $148m.
Is this a niche market? The stated TAM in the company’s S1 is $37bn, which we will come back to it in a moment

What are the metrics telling us at a high level? It seems to tick most if not all the boxes we would want for a Hypergrowth candidate, and the metrics are heading in the right direction. More in a second, but 121% DBNRR, 72% YoY Revenue Growth, 18.7% sequential Revenue Growth, 11% sequential customer base growth, 66% RPO increase YoY, 71% Gross Margin and stable, $310m cash on hand despite $17m in listing fees, not yet FCF but heading in the right direction with just -5% FCF margin in Q3 2021 despite listing fees

Is this not a highly competed space? While at first sight it might look like it competes with the likes of Google Analytics, I’d argue (and the CEO, Spenser Skates, always does) they play in a different segment. Products such as GAnalytics are really marketing-focused analytics products meant to drive different insights from the product-focussed Amplitude, whose core focus is to answer questions such as “How do I improve start trial to complete trial rate?” or “Where are users dropping off before subscription?”. In fact, to the extent tools such as GAnalytics are used for these purposes, I’d argue it is a case for being optimistic about Amplitude as it demonstrates a real need to drive these insights and customers have had to make do with inferior products given their need is so strong. This is why it is ranked #1 in Product Analytics on G2.

Is this not competing with the likes of Snowflake? Quite the opposite, they recently announced a partnership enabling a faster onboarding experience enabling Snowflake customers to leverage existing product data and derive insights on the Amplitude platform. This addresses one of the largest challenges Amplitude faces, i.e. the onboarding stage which takes 4 to 6 months with clients, so this will hopefully enable it to attract customers faster, an area to clearly monitor in next quarters. Interestingly, one of the most common use cases for Snowflake usage seems to be to store product data, according to the Amplitude CEO (Q3 2021 Earnings Call Q&A)

Do they even have a moat? It seems Amplitude enjoys two large benefits in terms of defensibility: first, the other side of the coin to their large customer acquisition process, is that customer switching costs are very high. Once implemented (~2-3 months from contract signing) it takes significant effort to move to a competitor as the platform becomes embedded in internal processes and the code for generating insights is embedded in the fabric of the customer organisation. Secondly, the product gets stickier with use: as customer interaction data is generated it gets increasingly harder to switch as you would lose out the ability to tap these insights and you’d have to start again. The question at this stage is how the Snowflake partnership might weaken this moat, something that will hopefully be addressed in coming Analyst Q&As as the partnership takes hold.

What’s unique about them? Their edge is two-fold: firstly, from a technical perspective, they seem to have developed a proprietary Behavioural Graph enabling flexible AND fast interaction with the data, something previously not doable with existing paradigms (for the techies, an SQL approach is very flexible as you can write any query but rather slow, while pre-computing insights is fast but not flexible as the moment you deviate from cached queries you hit the database speeds). Secondly, they seem to just understand customers much better than competitors and have built the industry-leading product, which is ironic given it is their very purpose to exist. To this point, it is important to consider the founding story, i.e. the three co-founders run a voice-recognition startup for 2 years before it failed and in the process they realised they lacked the tools to make good development decisions, so their existence is borne out of need, always a positive as it ensures you deeply understand your customers, vs stumbling upon a product-market fit without learning how to replicate that process for new products. This should give us confidence for new product developments in the pipeline

Who said new products? Up to now we have described the core “Amplitude Analytics” product, but recently the company released two new products in its all-encompassing Digital Optimisation System: Experiment. And Recommend. With Experiment, customers are able to run A/B tests from within the Amplitude platform to test new features and truly get to the holy-grail of data-driven product decisions. Interestingly, in the Q3 Earnings call, CEO Skates said “I think, that’s been one that’s been requested by our customer base for many, many years at this point” and commented positively on its early adoption despite Q3 being the first full quarter of general availability. Recommend on the other hand enables product teams to developed personalised experiences for each user based on their characteristics, so as to leverage their insights of what customer journeys lead to desirable results and better optimise each user’s experience. Clearly the products interact tightly between them and should act as a flywheel to drive up-selling, cross-selling and increasing DBNRR above 120%, but the next few quarters will tell this story better.

So what’s the vision here? Spenser Skates, CEO, is very clear what they are after: they want to be for the Chief Product Officer what Salesforce has been for the Chief Commercial/Sales Officer and Adobe has been for the Chief Marketing Officer. Interestingly, the CPO figure is itself not a very common one but it is a function that has increasingly been making its way in the C-Suite, demonstrating the increasing importance placed on this core competency. Gone are the days of throwing salespeople at a mediocre product and getting commercial traction with that, the new paradigm is frictionless optimal customer experiences, and Amplitude both rides and enabled this trend

Why did they not even IPO? CEO Spenser Skates is remarkably open with several Reddit Ask Me Anything (AMA) threads which are highly informative to read and give a flavour for the development of the company. When asked about direct listing vs IPO, Spenser said he couldn’t sleep knowing he’d be letting his shareholders leave 30/40% of their worth on the table given the well-document and well-understood IPO under-pricing for IBank’s customers, preferring instead to raise 150M in a Series F round shortly before the listing to anchor the valuation and raise sufficient cash to progress on their growth. With +300m in cash on the BS as of Q3 the company is well capitalised to pursue its growth and tells the story of a highly methodical CEO taking fiduciary duty extremely seriously, giving confidence (to me at least) in his execution. Also, many of his interviews paint the picture of a very modest, down to heart execution-focused person, as opposed to a story-teller with a large ego. Just saying.

Diving in the numbers:

Revenue


2019	 	 	$18.20	$20.10
2020	$22.30	$23.70	$26.40	$30.10
2021	$33.10	$39.30	$45.50

YoY Growth


	Q1	Q2	Q3	Q4
2019	 	 	 	 	 
2020	 	 	45.1%	49.8%
2021	48.4%	65.8%	72.3%

Sequentially (QoQ)


	Q1	Q2	Q3	Q4
2019	 	 	 	10.4%
2020	10.9%	6.3%	11.4%	14.0%
2021	10.0%	18.7%

These numbers seem to indicate an acceleration, however revenue guidance for Q4 2021 came in at $46.5m at the midpoint, which would represent a slowdown to 2% sequentially and 55% YoY. The only quarterly beat we have since public was of 4.6%, so assuming that’s a sandbag it would take us to $48.6m for 6.9% sequential and 61.6% YoY. When asked in the Q&A, the CFO attributed the cautious guidance to being a newly public company and wanting to be prudent with expectations. Also, he mentioned that the previous quarter’s strong guidance upon the direct-listing came with only a few weeks on the quarter left before reporting Q2 results, at a point in which they had great visibility on basically a full quarter of targets met. Seems reasonable enough of an explanation, and all other metrics do not tell the story of a slowing business so expectations are definitely higher than the guidance at this stage

Customers


	Q1	Q2	Q3	Q4	YR
2019				 739 	 739 
2020		 845 	 920 	 1,039 	 1,039 
2021		 1,280 	 1,417 		 
 

YoY


	Q1	Q2	Q3	Q4
2019	 	 	 	 
2020	 	 	 	40.6%
2021	 	51.5%	54.0%	

With only a few quarters of data it will take at least two more quarters to fully appreciate any potential seasonality with Q1 and Q4 comps; having said that, there indeed seems to be an acceleration in customer wins, so another positive signal. If they could maintain in Q4 2021 the same pace logo wins as Q3, that would take them 1570 customers for a 2021 YoY 51% increase vs 40.6% in 2020, again an acceleration

DBNRR

 
	Q1	Q2	Q3	Q4	YR
2019	 	 	 	116%	116%
2020	 	118%	119%	119%	119%
2021	 	119%	121%	 	 

Solidly hovering around 120%, the stated target and expected level that management guides for. This is not as high as some of the SaaS comps in the 140% range, but it is acceptable and healty given the stage of the company. There seems to be plenty of space for this to increase in the future as Amplitude rolls out two new products and can leverage its customer base for strong up-selling and cross-selling. (more in a second)

Operating Expenses
Limited quarterly data available here, but they seem fairly stable with S&M at 45% of revenue, R&D at 19% and G&A at 14%.

FCF Margins
Q3 2021 came in at a sizable -35% but was mostly driven by a sizable 16m of direct listing expenses hitting the P&L as a one-off, adjusting for which leaves the FCF margin at ~-5%.

This is coherent with a story of improving operating leverage, with FCF margins heading in the right direction:


2019	-24.4%
2020	-12.3%
H1 2021	-10.0%

It should come as no surprise, then, that the market thinks highly of Amplitude: currently trading at 28.6x NTM Revenue, it enjoys a similarly rich valuation to MNDY (28.2x) and SHOP (30.6x) who have longer track records of sustained performance

Key risks

  • New product risk, i.e. how successful will they be at replicating this early success in the core Experiment product in further adjacencies
  • How big their TAM really is, they claim it to be 37bn and not limited to software clients (and 26 FTSE 100 customers give us confidence) but it remains to be seen how effective AMPL will be at penetrating later-stage customers on the S-curve
  • Ability to sustain high-growth mode, but that is a common risk for hyper-growth companies, AMPL just has more to show given the lack of track record, so the next few customers will be very important

For me it’s eyes on next quarter with a 5% long position and looking to add more if these number keep shape, but curios what other members of the board think of this company?

YComp.

44 Likes

@YCompounder, thanks for that great summary.

My initial take is that Amplitude is kind of like an Alteryx but setup for a specific purpose, but running in the cloud as a service. A key advantage/differentiator for both Alteryx and Amplitude is their user interface, which enables non-technies to run analytics. Amplitude has the advantage of natively running in the cloud, so it’s provided as a service whereas Alteryx to date has required an on-premise Windows machine(s).

Which brings up a question: Assuming Amplitude is great at what it does, how limiting is it that it’s good for are relatively narrow set of purposes (user centric product analysis). Their internal database is setup specifically for fast user-based database joins, which might preclude other uses of the data collected.

Now, if customers import data they collected on their own into Amplitude (from the newly announced Snowflake connector, for instance), then this wouldn’t be a problem as those customers have their data in Snowflake and can do anything they want with it. But, customers that have instrumented their code with Amplitude’s APIs have that data solely within Amplitude and thus would be restricted to whatever Amplitude provides.

Maybe this isn’t a big deal. Maybe the use cases supported by Amplitude are strong enough and the ROI is high enough that it doesn’t matter. Maybe there really aren’t other important uses for that data. Maybe this is an expansion opportunity for Amplitude down the road. Maybe Amplitude provides a data export feature.

I’m not saying Amplitude is a bad investment, I’m just curious.

30 Likes