Kellblog

A Simple Rule For When To Consider Selling Your Startup

Posted on October 30, 2024 | 1 comment

The cleverest answer I’ve heard to the question “When would you sell your startup?” is, “When somebody offers me more than I think it’s worth.”

It’s clever alright, but it’s not that helpful. It’s a meta-answer that sidesteps the question of value. In this post, I’m going to offer a simple rule for when you should consider selling your company if someone comes along and makes a serious offer.

Selling a company is a hugely difficult decision that involves both personal considerations and big, strategic questions like:

If I say no, what might that strategic suitor do instead? Will my top partner become my top competitor, potentially overnight?
What is happening to the space at large? Are my competitors being gobbled up? Might I be the one left without a chair when the category consolidation music stops?
Does the market require a Switzerland, an independent vendor who’s not owned by any of the megavendors? (For example, as data integration has historically.)
If I “keep on keeping on” — given my size, growth rate, and financing ability — where I am likely to end up? As a clear market leader? As an undifferentiated, fifth-place also-ran? (And those don’t trade for median multiples.)
Or, will I “pull a VMware” and sell a business for $625M that will one day be worth $60B? (And conversely, if I own 30% of it, how much difference will that 100x uplift actually make in my life?)

In this post, we’re going to keep it simple by focusing not on whether you should sell your company, but on whether you should consider selling it.

And, as is often the case, I have a simple rule. You should consider selling your company when an offer takes three years of risk off the table.

What does that mean?

Go look at your three-year model (and this is one of many reasons why you should have one)
Find your ARR is twelve quarters out
Multiply that ARR by what you think will be a reasonable ARR multiple given your future growth rate
If the offer on the table is greater than or equal to the number you just calculated, you should consider selling.

Why do I think this formula works? Because:

Most three-year models are optimistic. For most companies, you’re looking at a best-case estimate of what ARR will be in three years.
You’ll probably overestimate the “reasonable” multiple. If the market data suggests 4-6x, you might round up to 6-8x. Human nature.
You’ll therefore generally arrive at a generous future valuation that will take no small amount of work to realize.
And a lot of shit can go wrong along the way.

Why consider taking that valuation? Simple. Because building companies is hard. As one founder often said, “not just harder than you think, but harder than you can possibly imagine.” I’ve played a leading part in building four enterprise software companies and I’d say it’s just plain hard. But maybe that’s because I have a better imagination. Or an imagination fueled by more experience.

So, if you want to pay me now what I think the company will objectively be worth in three years — if everything goes smashingly — then I’m going to need to seriously consider that offer.

Note that I am not suggesting you should automatically take any offer that’s N times your 3-year forward ARR (for whatever value of N). You need to answer all those personal and strategic questions first. But I am also saying that — unless conversely you see storm clouds on the horizon — that you shouldn’t invest too much time considering offers that take 1-2 years of operating risk off the table.

Why?

Your one-year-forward ARR is basically your operating plan. And, if you’ve followed my planning advice (“make a plan that you can beat“), then you should have pretty high confidence in that plan. So an offer that takes one year of risk off the table shouldn’t be that compelling.
Your two-year-forward ARR is not the layup that your one-year plan should be, but if you think your model is realistic — and I admit this is completely subjective — selling off two-year-forward ARR just doesn’t seem worth it. You’re trading away future upside for a number that you’re still pretty confident you can hit. Two years is a long time, yes, but not that long.

For me, at three years, the tone changes. A lot can happen in three years. New competitors. Category consolidation. New vendors entering the space. Bad C-level hires. Product development disasters. Failed acquisitions. Geographic restarts. Channel programs that don’t take. And other scary things that go bump in the corporate night.

Yes, everything may go right over your coming three years. Maybe you’ll even beat that optimistic three-year model.

But at three years, I start to do some hard thinking about both strategic and operational risks. I’ll need to feel very good about the future to say, “no thanks, we’ll roll the dice.” Particularly if that suitor is a megavendor with intent to enter the market anyway. Or, if multiples are currently high relative to historic averages — meaning that I might get offered 10x $50M today, but by the time I’m actually at $50M, the market may be trading at 6x. That’s three more years of work for 40% less money assuming everything goes great. All because multiples moved down on me.

Let’s insert a little model to make this concrete:

Using my rule, if someone offered me $675M for this company, I’d have to seriously consider it. Note that’s 45x this year’s ending ARR and 22x next year’s. But since it’s November, I’m already worth $233M, the 15x multiple coming courtesy of my 107% growth rate. (Feel free to quibble with me on the numbers; I think they’re representative, but the approach is the point.)

If someone offered me $354M, then I’d say no and roll the dice on achieving my operating plan. Because once I do, I’ll be objectively worth that in 12 months.
At $535M, I start to get queasy because that’s 2.3x what the company is objectively worth today. This is why I made the model — to make things concrete — because I might well consider that offer, particularly because my decelerating growth drops my 2027 multiple to 8x, meaning only $141M of incremental value is created in 2027. I wouldn’t definitely consider it, but I probably would and I’d be arguing the whole time that we can make it worth more given “only” three years’ work. (My change in tune here is a negotiation posture.)
At $676M, I’d definitely consider the offer for the reason stated above: that’s what the company should be objectively worth in three years if everything goes well. And a lot can go not-so-well over a three-year period.

Note that I did one sneaky thing in the model. I included the number of AEs I’d need to make those numbers using that as a rough proxy for work. To make the plan work, I’m going to have to hire 35 reps — net of attrition — over the next three years. And all the support resources they need and/or generate, e.g., SDRs, SCs, managers, post-sales consultants, CSMs. For me, it helps to make the anticipated work visceral.

Let me address some anticipated objections to this approach:

It doesn’t consider long-term strategic value. You’re right. It doesn’t. That’s why you need to consider that under the big strategic questions. Don’t pull a VMware. Or, arguably, a YouTube.
If I did this three years ago, I’d have sold for pennies. To be concrete, let’s say the company’s annual ARR ramp was (0, 1.0, 2.5, 7.5, 15.5), which dovetails into the table above. That means you’d have seriously considered an offer of $46.5M three years ago when you were $1.0M in ARR. If you’re VC backed, there’s no way you’d sell, so you can consider it as much as you want. And if you looked at personal and strategic considerations, you probably would have said no anyway. But yes, the rule doesn’t scale particularly well back to $0.
This will backfire going into a valuation bubble. And it will. Say multiples now are 6-8x for your growth rate and you model off 6-8x in your three-year calculation. If the market gets frothy, those multiples could double to 12-16x. Ergo, this formula will underestimate your future value by half. I have two responses: (1) the opposite is true as well; you win when you apply bubble multiples to future non-bubble ones, and (2) the ARR figure is probably over-estimated which should mitigate but not eliminate that effect.
It doesn’t include a hockey stick when we hit some market inflection point. That’s Silicon Valley speak for the model might underestimate three-year ARR because of [insert miracle here]. And it might. But for every 100 companies I see waiting for those miracles, maybe 2-4 get them. It’s rare. And if you really think that inflection point is going to happen, put it in the three-year model.
All the great founders are all-in, YOLO. I have two words for you: survivor bias. I know lots of great founders who were all-in, got sunk on the river, and wish they’d taken some money off the table. Ultimately, this will come down to your personal goals. (And it’s one reason why I think VCs love second-time founders. Dave Duffield was never going to sell Workday too early because he had all the money he ever needed from PeopleSoft. And since VCs would generally rather sell too late than too early — because it’s a “hits business” — that creates a deep, natural alignment.)

My purpose here was to give you a rational framework for thinking about this decision, and here it is: think about how many years of risk gets taken off the table. That’s what you get in exchange for trading away whatever potentially bright future awaits.

The rest is up to you — and your board. Good luck with it.

# # #

(Revised 3:59pm. Sorry, this was accidentally published before final spellchecking and copy editing was complete.)

1 Comment

Posted in CEO, M&A, Silicon Valley, Startups, VC

A Box of Rain Will Ease the Pain

Posted on October 27, 2024 | 2 comments

While I’d been using the Internet since the early 1980s in my student job at Lawrence Berkeley Lab, the first time I remember using the worldwide web was in the mid 1990s. Well, August 9th, 1995, to be more specific.

I’d recently moved to Paris and heard about this layer atop the Internet that relied on hypertext transfer protocol to connect web browsers and web sites. Having seen what Apple had done with hypertext up that point, I wasn’t prepared to be impressed. So I fired up a browser and, to reconnect with home, I went to sfchronicle.com. The headline read:

Jerry Garcia Dies

“God, I hate this thing!” As an inveterate deadhead, the news was devastating if not entirely surprising. The episode set my web adoption back by a few years. While I won’t dive into my history following the band, I’ll show you the back of the car that we keep at our house in Oregon. (Morning Dew.)

Four years earlier, we’d lost the power in our Marin County home the night Bill Graham died in a helicopter crash west of Vallejo. If I was connected to Graham’s death via a power line, I was connected to Garcia’s via the web.

An SMS message connected me to last Friday’s death of Phil Lesh. Sent by a friend from so long ago that my phone no longer recognized his number, the news arrived as anonymous text message, containing only a link to the story.

I’ve written before (and as recently as three weeks ago) about business lessons from the Dead, so I won’t cover that again. Instead, as my tribute to Phil, I’ll write briefly about the power of metaphor using one of the relatively few Dead songs he wrote: Box of Rain. It’s also one of my favorites.

The song deals with Lesh’s feelings during the lingering, terminal illness faced by his father. Grateful Dead lyricist Robert Hunter worked with Lesh on the lyrics, and they are some of Hunter’s finest.

You can hear the frustration and impotence in battling illness via lines like:

“What do you you want me to do, to do for you, to see you through?”

The power and beauty of the song, however, comes from the primary metaphor: the box of rain.

“Just a box of rain, wind, and water. Believe it if you need it, if you don’t just pass it on.”

But what is this box of rain? It’s a metaphor. In fact, it’s a metaphor within a metaphor.

Hunter was thinking along the lines of a “ball of rain,” but that was probably both too obvious and insufficiently poetic. So the ball became a box. (Hence, the inner metaphor.)

So what is this ball of rain, wind, and water?

It depends on perspective, and in this case you’re going to need a wide one. Seen from far enough away, that ball of rain is our home. The earth.

“It’s just a box of rain, I don’t know who put it there. Believe it if you need it, and leave it if you dare.”

Once you understand the metaphor, that’s pretty literal.

“And it’s just a box of rain, or a ribbon for your hair.”

Hunter loved to write about certain things, such as calliopes and ribbons. While hard to interpret, I think this line is another metaphor. What do ribbons do for hair? Make it more beautiful. What does the box of rain do for the universe? The same thing. The earth is just a ribbon in the hair of the universe.

It adds a sense of utter smallness, rivaled only by how Tralfamadore used the earth in The Sirens of Titan.

(I feel obliged to say that Hunter hated to interpret his lyrics, maintaining that it wasn’t about what the words meant to him: it was about what they meant to you. He didn’t want his meaning to ruin your meaning. At first, the puzzle-solver in me found this offensive, but over time I’ve come to realize that it’s actually pretty cool.)

Now we’re ready for the last line.

“Such a long, long time to be gone and a short time to be there.”

There, of course, being here. On our box of rain. In this life. On this earth.

Thank you Phil for always putting the music first, the culture, and the reminder. May we all spend our time here as well as you spent yours.

2 Comments

Posted in Grateful Dead

The Impact of AI on SaaS Metrics: Video Now Available

Posted on October 24, 2024 | 1 comment

Just a quick post to highlight that the good people of Benchmarkit, host of SaaS Metrics Palooza 24, have posted the video of my presentation, The Impact of AI on SaaS metrics. The slides are here.

1 Comment

Posted in AI, Metrics, SaaS, Startups, VC

Design Your Organization for the Conflicts You Want to Hear About

Posted on October 12, 2024 | 2 comments

Organization design seems a popular topic these days. Maybe it’s the downturn. Maybe it’s just planning season. But either way, many people are asking me questions about how to design their organizations for 2025 and beyond. Questions like:

Should marketing report into sales?
Should engineering and product management (PM) report into a combined product org?
Should we unite customer success and sales?
Should North American and Europe report into a single head of sales?

The argument for combining teams is always about reducing span of control. This is a goal that many CEOs (and some boards) share, but one that somehow escaped one of the world’s most successful entrepreneurs, Jensen Huang, who has about 60 direct reports.

While 60 seems a bit much, I’ve frankly never understood span-of-control reduction as a top organization design goal. As CEO, you should be managing senior people so they shouldn’t take that much time. So, why not have 8 or 10 direct reports? If you can’t handle that, maybe the problem isn’t that you have too many reports, but that you’re managing them too closely. Maybe the solution isn’t to reduce their number, but to loosen the reins.

I have two rules for organization design:

Design for conflict. Specifically, design your organization for the conflicts you want to hear about.
Ensure value-add. Don’t put thing B under thing A unless the executive in charge can add value to both.

Design for Conflict

When you put, for example, engineering under product, what don’t you hear about anymore? Conflicts between PM and ENG about the time and resources required to build things. Those conflicts get silenced because the SVP of Product will suppress them, resolving them in the family.

When you put marketing under sales, what don’t you hear about anymore? Conflicts about whether sales strategy is too unfocused to enable marketing targeting. Or whether sales follows-up on new oppties in a timely manner. Those get silenced because the CRO wants to manage their own house. “Let’s resolve that at the sales QBR, not the e-staff meeting.”

When you put customer success under sales, what don’t you hear about anymore? Conflicts about whether sales is overselling to the point that customers won’t be successful and ergo won’t renew. If churn looks high, well, it must be the product. It’s not delivering, but against what expectations, set by whom? All silenced.

The rule is simple. By combining two departments, you are asking one person to resolve the conflicts between them. They’re not evil to do so; it’s the job you asked them to do. They will keep these conflicts in the family. And, as the organization grows, you will hire increasingly senior people to do just that. But with each layer and with each combination, you get more insulated from the ground truth.

So the question is simple: which conflicts do you want to hear about? Which do you want to pay someone else to resolve and which do you want brought to your office? Which are strategic to the company and potentially involve Crux-level issues?

If you separate PM and ENG, you’ll hear a lot about specs, resources, and timelines.
If you separate sales and marketing, you’ll hear a lot about awareness, leads, and follow-up.
If you separate customer success and sales, you’ll hear a lot about over-selling.

There’s no magical answer here. Just a framework for thinking about it. Determine the conflicts you want to hear about — presumably because you can add the most value in resolving them — and then design the organization to make sure you do.

Ensure Value-Add

The other principle is to always ensure value-add, beyond the (sometimes merely assumed) alignment that comes from having a common boss. So, sales wants the SDRs to report to them? Why? Has the sales VP managed an SDR team before? Are they good at it? Are they even interested in it? Can they add value? Are they sufficiently metrics and process-oriented, particularly if the VP comes from an enterprise background?

This principle drives a number of positive effects:

It defeats empire building. Sometimes the VP wants the SDR team not because they care about them, but because they want a bigger organization. Or they think it will look good on their resume for their next job search. They’re not actually interested in the job. They’re interested only in being able to say that they did it. That’s not good enough.
It encourages learning and development. When the VP of sales first asks about managing the SDRs, you can tell them to go make themselves a valid candidate. Get close to the SDRs now. Understand their challenges and offer to help out. Network with friends and colleagues on SDR team management. Read up on best practices. Convince me that you’d make the short list of candidates and then we can have a conversation.
You attract stronger department heads. Everyone should work for someone they can learn from. Saying the boss is the boss because, “well, we had to plug the team in somewhere,” is a terrible reason for an organizational structure. If you apply the value-add rule, functions will tend to report higher in the chain, creating a flatter org, and be placed only under those who can add value to them. This, in turn, attracts stronger candidates to run them. Who wants to be the CMO when it reports to a CRO who understands nothing about marketing? Nobody.

One great example is whether the VP of European Sales should report to the existing VP of Sales when you expand into Europe. If your VP of Sales is clever, they’ve already given themselves the title “VP of Worldwide Sales,” and you let them do it because it was moot at the time. But now they’ll argue it’s a demotion if Europe doesn’t report to them. And they’ll argue that they know how to sell the software in North America (really, the USA) and that knowledge should translate anywhere. And that everybody does it this way. You can almost hear them screaming: pick me, pick me!

But what they should be screaming is: I can add value, I can add value! And if they can, you should listen. But my questions would be:

Do you have a passport? (This wipes out about half of Americans.)
Have you ever lived in Europe?
Do you speak any European languages?
Have you ever sold and/or managed people in Europe?
Do you you have a network of people we can hire in Europe?
Do you have relationships with contacts at target customers in Europe?
Do you know any strategic partners we can work with in Europe?

So, other than not having a passport, never having been there, knowing no one, and not being able to communicate, you strike me as an outstanding candidate for the job.

We do this all the time nevertheless, and Europeans have grown accustomed to reporting into people who can’t add value. But for my nickel, I’d rather hire a VP of EMEA who had great answers to my questions and reported directly to me.

Mitigation Strategies

As your organization grows, you will invariably combine teams and lose your line of communication into certain conflicts. I know three ways to mitigate this:

Build a culture of transparency where direct reports into e-staffers are encouraged to and rewarded for speaking frankly about in-the-family problems.
Run an extended QBR. Don’t just invite the e-staff to the quarterly business review, but also invite people among their direct reports. For example, the head of customer success if it reports into the CRO, or the head of engineering if it reports into product. Ask them to deliver the same, standard presentation that the e-staffers do. This effectively flattens the organization by creating an extended leadership team that goes beyond the CEO’s direct reports.
Use reporting. Good reporting can reach through organizational layers and keep you in touch with what’s happening. For example, even if customer support doesn’t report to you and isn’t represented on the extended leadership team, you can still keep an eye on metrics and KPIs as well as simply on OKRs.

In this post, I’ve argued that the primary goal in organization design should not be reducing of span-of-control, but in surfacing conflicts most important to the company. I’ve also introduced a value-add rule that says no department should report into an executive who can’t add value to it. And finally, knowing that consolidation is inevitable over time as a successful company scales, I’ve offered three strategies to mitigate some of the signal loss that comes with such expansion.

2 Comments

Posted in Entrepreneurship, Leadership, Management, Scaling, Startups

Tagged Conflict, Organization

Slides from SaaS Metrics Palooza 2024: The Impact of AI on SaaS Metrics

Posted on October 8, 2024 | 1 comment

Just a quick post to share my slides from today’s presentation at SaaS Metrics Palooza 2024, entitled The Impact of AI on SaaS Metrics.

The short summary is:

The concept of ARR is already challenged by monthly-varying pricing, e.g., usage-based pricing.
AI will exacerbate that problem, bringing new forms of value-based pricing, e.g, unit-of-work or outcome-based pricing.
There are two schools of thought on dealing with this: (1) split the ARR baby into baseline and variable, then analyze the baseline as if nothing has changed, and (2) spend is truth, where we substitute trailing spend for ARR. I’m in the second camp.
AI will, gasp, require us to think about cost, something we don’t really like to do in the software business and something we’ve historically been able to kind of ignore.
All the heavy lifting is going to move to the pricing model.

In short, to know ARR we used to read contracts. In the future, we’re going to read invoices, instead.

Yes, for internal reporting we will do a lot of pricing model analysis and examination of the base/variable split. But for external reporting, the big six SaaS metrics all depend on ARR and going forward that won’t change. We’ll just use some proxy for ARR, as many quietly do already today.

Like a duck, nothing will change much on the surface, but they’ll be a lot of activity underneath. And the metrics won’t mean quite the same thing as they once did. For example, ARR and NRR will become less forward looking and work less well as leading indicators.