Product-led growth is overrated

(for vertical startups but also maybe for every startup)

Hey folks,

If last week was a strategy and general mindset piece around vSaaS GTM, this is more a note on whether it is worth pursuing product-led growth as a vSaaS business… or really at all. No doubt, founders are capable of parsing this out themselves. This piece is an excuse to go on a bit of an exploration around product-led growth’s prospects in the 2020s.

Frankly, I’m skeptical. Silicon Valley is ready to go all in on product led growth. People are really impressed (rightfully so) by the financial metrics of PLG businesses. Some of this is legitimate, some of it is insane (I think putting Snowflake as a PLG business is ridiculous). But generally the story goes, that business acquire customers more cost effectively, with a higher net dollar retention rate year over year.

Leaving aside whether the businesses that get cited are actually evidence of successful PLG motions or are more traditional sales motions with a land and expand model (which is totally what Snowflake is)¹, every company is looking to capitalize on the PLG frenzy. But I think, that’s what it is: a frenzy. There’s good reason to believe that PLG was a result of anomalous conditions or is grossly overstated in impact. But more than that, I just don’t think it works very well for the types of businesses that vSaaS founders will build.

Founders must make a decision around their specific GTM based upon the unique circumstances in your vertical. And my hunch is that this will look like a high touch process.

So today is two-fold: a more extensive look at product-led growth and then an abbreviated look at the virtues of a more traditional sales process.

Product-led growth

The devil is always in the details and as much as grand theorizing matters for figuring out how vertical SaaS companies should relate to the current empires like last week’s piece, it matters a lot more how you think about the specifics of GTM as a vertical SaaS company.

In the midst of a recession and most definitely a bear market for technology investment, product-led growth is viewed as a panacea for growth pains at startups.

Tomasz Tunguz paints the motion this way:

PLG/PLS motions typically couple three things:

1. Free-trial/freemium product to attract a large top of funnel (TOFU)

2. Product analytics to monitor prospects through their customer journey in the middle of the funnel (MOFU), and predict the best time to contact the user to sell them, at which point they are product-qualified leads (PQL).

3. Sales-assisted close to maximize the conversion rate (yep, you guessed it - BOFU).

In many types of companies, this can work extremely well and start an incredible flywheel. Acquire end users at a low cost, these users evangelize your product, you acquire more users at even lower costs. Eventually you can upsell companies to an enterprise package and realize revenue from the motion.

In the abstract, this is capital efficient. With a viral and accessible product, your marketing spend compounds growth. Each user acquired with ad dollars should bring in additional users. In theory, as you start to target enterprise deals, your cost of sales should also be far more efficient with resources only going to the accounts that have product usage and thus that are more likely to pay for an enterprise package.

Starting this flywheel then becomes a key goal.² And given the success of this strategy in the 2010s with notable companies like Twilio now public, it's no surprise that product led growth is becoming buzzy.

Tomasz has even predicted that it will become the standard GTM motion.³ After all, one solution for growth in a world where capital efficiency matters and growth is costly is to make a product that can effectively grow itself.

To sum it up I think that PLG rests on some foundations that may no longer hold true.

Capital Efficiency

The entire notion of product-led growth as a successful strategy relies upon the assumption you can get the flywheel going in a cost effective manner. If you are simply burning unprecedented amounts of money to acquire non-paying users, you haven’t built a flywheel, you have built a money pit.

So my question is really: Is capital efficiency a result of the PLG motion itself or an arbitrage upon anomalous conditions that no longer hold true?

To put it bluntly, Tim Cook might be a harbinger of the end of tailored advertising and with it a far harder path to a flywheel.

For those unfamiliar, in 2021, Apple decided it was time to tackle user privacy and reduce the ability of companies to track user data.⁴ It used to be the case that companies could do things like retarget users that came upon their site, get precise data into which of their ads were working and which weren’t, and much more. This was a huge boon to companies who primarily acquired users through the internet. DTC brands popped up everywhere and consumer businesses could acquire customers cost effectively.

After 2021 and iOS 14, this is no longer true for much of the consumer world. B2C has been hard pressed to acquire users cost effectively. Customer acquisition costs are high and a lot of performance marketing simply doesn’t work on traditional methods.

Is the fallout limited to consumer? Maybe. However, Google has been working on a privacy focused framework that may also hinder much of the B2B space.

This will inevitably raise cost per lead. But less targeted data is only half the story. Far more pressing is how many ad dollars are going after the same user group.

Startups killed PLG

The entire marketing world agrees with me on the first point: it’s way harder to acquire users or to predict which ads are working. On this next point, they agree less so! But startups are killing PLG.

Here’s what I mean: PLG is becoming priced in. As investors continue to invest in companies that invest in PLG, a couple things will happen: a) more companies adopt the same marketing strategies which includes investing more in ads, content marketing, and b) more companies will have freemium tiers.

The problem is that while ad spend is relatively unbounded and you can always tack on a free trial or new freemium tier, end users are bounded.⁵ What this means is that there can be an ever increasing amount of ad dollars chasing the same group of end users.

And it’s not just that end users are bounded. If end users could use every single software product, this wouldn’t really matter. But they can’t. And they won’t. End users are finite. And their attention even more so. There are only so many apps or developer tools that an end user can… use.

After all, how many productivity apps can you realistically try? How many dev tools?More importantly how do you make sense of the assault of ads and free trials when you try to find a new tool for your use case?

If this is a product designed with making your individual productivity better, you may simply pick whatever product you happen upon that broadly meets your goals. You might try a couple but realistically you are going to settle for one of the top search terms.

Are you going to go through all 42? Or click on the ones paying for your attention?

So the impetus then becomes on companies to spend their way to the top of searches and generate good content for SEO. And if you don’t? Good luck ever achieving PLG.

Let’s say that the product is business critical, your end user is going to do a bit more research. They will spend time looking through the options, read through docs, and spend time in a free trial. But they won’t stop there. After all, savvy end users making decisions that will affect their company substantially are not going to stop with publicly available information. Instead… they are going to talk to sales to get non-public information.⁶ The more companies that pursue a PLG motion, the more critical this becomes for end users.⁷

In other words, just because you gave the end user product accessibility doesn’t mean you are growing with the product. In a world where everyone’s product is available to try, it doesn’t mean you win on that alone.

So in the case where the product is business critical, you talk to sales. And in the case where your product isn’t, it’s a death spiral of ad spend where your end user has very little costs to move onto the next product in the same category. After all, they will give your end user a freemium version too! In order to prevent this, you probably end up trying to go win a buyer in the company right away with… you guessed it: a sales team. Again, it’s not clear that this is a de facto capital efficient strategy.

So the story of the 2020s may be one where see the product-led motion fail or at least be far more dependent upon sales and marketing than the network effects thesis would lead you to believe.

Capital Efficiency kills PLG

Lastly, there’s a solid chance that the very notion of capital efficiency as a priority kills a lot of PLG motions. When capital was nearly unlimited, it was quite easy for employees to find great products, start using them, and then get a manager to justify the spend on an upgraded package. But we left a bull market. While software spend won't be reduced, you need to prove out that your software is worth the spend. And even if it is, your conversion rate will be lower. And if you don’t? Good luck convincing other companies focused on capital efficiency why they should buy your product.

So we reach the ultimate question: If you have a an ever-rising cost per lead to generate product qualified leads, an additional cost of sales with a sales team, and a lower conversion rate, did you really grow capital efficiently? Not really. My hunch is that we may not see the network effects of PLG pay off in actual revenue with many different software businesses.

Regardless of whether I am right about PLG’s prospects generally, I am convinced that for vSaaS, it will be of limited value.

VSaaS Concerns

All of my concerns about PLG for vSaaS companies can be summed up in one sentence: You are most likely working in an industry chronically under-penetrated by software. Your goal will be as much to convince them of the value of software for their workflow as much as it is your company. Second, your end user has far more pressing things to do than figure out the best software stack.

End User Makeup

PLG craves a certain type of end user. Where PLG has worked, the end user has certain traits. First, the end user has agency around the tools they use. Companies want their developers to be happy about their software stack and thus let them pick the right tooling. Importantly, not only do these end users have agency, they also have the desire to pick their tools.

When you have a high agency end-user with the desire to pick their tools, it pays off to allow them to easily experience your product. Intuitively, this means that PLG works best in contexts where the end user understands and appreciates the value of software. And right there lies your problem! As a vSaaS, chances are your end user isn’t seeking out software.

So contrary to other startups, where their PLG is targeted at developers and more technologically sophisticated knowledge workers, yours will most likely be aimed at workers who are entirely unconvinced of the value of technology and in some industries may even see it as a way to automate them out of their job. And that’s assuming that your end user has the agency to even make a decision around the software stack. In industries characterized by small businesses, there’s a fat chance. A freemium tier isn’t persuading anyone.

So the first question you must answer: do my end users have the agency and the desire to pick their tools? To the extent your answer is an immediate no, forsake PLG. And even if the answer is yes, you run into a second problem.

Implementation

Another unstated rule of PLG is that cost of implementation must be close to zero. Users must be able to go in and use the product immediately. Usually this is because PLG companies can plug into preexisting infrastructure at a company through a seamless integration or the devs will just do it themselves.

For most vSaaS companies, there is going to be a gnarly implementation path. You can’t rely upon preexisting infrastructure with a seamless implementation nor on users who will do an implementation properly. The more meaningful the workflow, the more your company will have to support implementation. Every vSaaS company has discovered this and all involve customer support teams to help with the journey. ⁸

To sum it up, if you are working in an industry with low desire end users, and with hard implementations, it makes very little sense to pursue a PLG motion. There may be some exceptions. For instance, it might make sense if there is some preexisting infrastructure you wish to build upon or plug into. And PLG can work wonderfully as a way to expand the number of workflows you serve once you are initially in the business.

But by and large, PLG is not worth the energy nor the ad dollars. It won’t be capital efficient for most and it won’t be practical for vSaaS in the slightest.

But there’s good news: traditional sales and marketing still works and a good motion can be a boon to a business in reaching its end state: becoming an industry partner.

An industry partner

The ultimate goal of every vSaaS is to become the trusted software partner for their vertical. This is not an easy task. You become an industry partner by building world-class software and then systematically tearing down misconceptions that have penetrated the vertical about what software is capable of.

And at first, nobody will really believe you. In fact, if you ask them if they would use your product when you’re first starting out, you may run into the same challenge Peter Gassner of Veeva did: they simply won’t buy that your product is worth anything.

Watch a minute of this interview starting at 16:58:⁹

Veeva was built for customers that didn't really believe that the software was worth anything at the start. Fast forward to today, Veeva is the industry partner for the life sciences with over 50% of their vertical utilizing their software. So like Gassner, an initial lack of understanding shouldn’t deter you. It’s even okay if they don’t think your software is worth a damn. It’s simply an indicator that you really do need to focus on proving value as an industry partner.

And to do so will require not only a great product mind like Gassner, but someone who can go tear down the barriers like Matt Wallach.¹⁰ You need to find ways to prove the value of your software and this often won't be in a free trial, it will be in proving yourself out as a partner.

It will require deeply understanding your industry, chatting with your prospective clients regularly, and relentlessly demonstrating that you can fulfill needs that are currently unmet: often that they aren’t even aware of. So casts a vision of this partnership, even if you enter the vertical with only one workflow.

Burning the Boats

Investment in a partnership can have an extraordinary impact. In effect, the message you are trying to convey is that there is no alternate route for your success. It is only by winning their industry that you will win. In effect, you have burned the boats. There is no going back. You are here to win their business. This will focus you not only on creating an incredible product worthy of the boat burning, but also in cultivating the motion you must use to win their business long term. Some of this will depend on the size of the vertical and deals. But overall, your goal is to invest in a motion that creates trust, understanding, and on some level: an emotional connection. Tear down the misconceptions and the barriers and you won’t need ad spend. Your virality will come from your reputation as the industry partner. You will win the vertical.

Professional Services

One last component of becoming an industry partner and devoting yourself to a high touch process: partners have a mutual respect for the value each brings and are willing to compensate them accordingly. In many verticals where implementation is especially costly, this means that companies not only will compensate you for your software, but they will pay you for services rendered around implementing the software.

Your cost centers associated with implementation can become profit generators all the while it more deeply entrenches you in the industry as a trusted thought partner. Just make sure you are actually nailing the implementation.

In short, to the extent your ambition is to be an industry partner across the whole range of workflows, it pays to broadcast that from day 1. Capital efficient sales can be an asset and the post-sale implementation can also be a revenue driver. The key? Build for the industry and then make sure the industry knows it. These may be long sales cycles, but you can be confident that as you win them over, you are building a product and company that will not be easily replaced. True partners never are.

Next Up

Next week is the first Verticalized industry deep dive. I won’t announce it quite yet. The plan from here is to break into a monthly cadence where two weeks cover specific industries or companies and the other two weeks are more varied on a whole host of matters related to vertical SaaS. I’ve quite enjoyed writing these and it’s been fun to see increased readership so I am hoping to publish more frequently from here.

Cheers to the week ahead!

1. Yes, they do offer a free trial. I get that’s technically why. But their NDR is far more related to their consumption model than anything.

2. There’s a lot of great material on this and some of the principles associated with the strategy. I favor Bessemer’s guide here.

3. https://tomtunguz.com/2021-predictions/

4. Conveniently, this seems to be an absolute boon to Apple’s own ad platform.

5. Except when boards do in fact bound it.

6. Even developers who are the most savvy of savvy software buyers and who do not want to be sold to will probably not make a move on critical software without talking to the company.

7. This doesn’t mean that end users will want to talk to sales. In fact, they’ll express annoyance. But they will talk to sales, because they want various assurances that aren’t available simply by trying it.

8. Two implementation documents: Procore and Toast. Note that these are both mature companies with a lot of the implementation process fully figured out. This will not be the case for early products.

9. Substack doesn’t allow YouTube clips for some reason. Apologies for the friction.

10. Wallach was a former product guy in the life sciences who could drive the initial commercial efforts.