Why we're doing things that don't scale

http://37signals.com/svn/posts/3589-why-were-doing-things-that-dont-scale

by mh_ • 13 years ago

117 34 13 years ago

patio112• 13 years ago

I don't know if this is quite public knowledge yet, so: human-assisted onboarding is sort of catching on at a lot of SaaS companies in the B2B space, at price points which are a bit north of the top end of typical 37signals/etc apps. (Squarely in the zone for this new product, though.)

A lot of people find it very, very difficult to get started from a blank slate. Automated onboarding/email sequences/video tutorials/etc gets you to a point, but it still leaves guesswork in the process and (candidly) one can always blow off a computer but blowing off a person who has been talking to you for weeks is socially harder to do.

This also allows you to do basically a hybridized product/consulting model, since if (without loss of generality) Jason Fried is spending 2 hours for you talking about your business' challenges as he gets you up to speed on KYC, that's essentially equivalent from the customer's perspective to "Rent Jason's brains", but (crucially) the majority of the economic activity actually comes from the product whose use is enabled by the mini-consult rather than through consulting itself. This lets it scale to the moon. (Also, as every consulting company at scale will tell you, you can have Jason Fried's insights on your company actually delivered by an understudy. 37signals certainly doesn't lack for smart people who understand the 37signals playbook.)

This is currently giving me fits, since I haven't figured out a great way to deliver it consistently for non-enterprise accounts given that I'm 14 time zones away from many customers, but it's one of the big things I'm working on currently. The word on the B2B SaaS grapevine is it works very, very, very well at moving $X,000 to $X0,000 per year SaaS accounts.

I'll blog about this if I ever manage to systemize it successfully. (As usual, the plan is 1) Throw stuff at the wall and figure out what works. 2) Write down what worked and execute against that plan. 3) Pay someone to do #2. I'm currently somewhere in stage #1.)

hkarthik• 13 years ago

I have my reservations about this, because it feels like B2B SaaS is now going upmarket by reaching for the familiar Enterprise, high-touch sales process.

We've all been through this, and we know where it leads.

You have to bring on dedicated high-touch sales people to "scale" the business.

Sales people can rarely bring the same passion or insight as the original founders when approaching each lead. So they start doing more dastardly things (See Zed Shaw:http://zedshaw.com/essays/control_and_responsibility.html).

Going up market means compromising more on the scope of what the products do. As individual deals become more lucrative, the incentive to close them grows. As this incentive increases, the products often suffer and you get the Bananas in the Lasagna problem (See Jason Fried: http://entrepreneursunpluggd.com/blog/37signals-jason-fried-...).

So my question to the folks advocating this "new" approach is, how are you going to avoid all those things?

jt2190• 13 years ago

Watch Gail Goodman's presentation "How to negotiate the long, slow, SaaS ramp of death" [1] for an interesting discussion of this. (Skip ahead to 20:42 for her description of "coaches".)

[1] http://businessofsoftware.org/2013/02/gail-goodman-constant-...

(edit: presentation title)

patio112• 13 years ago

You have to bring on dedicated high-touch sales people to "scale" the business.

So it's very much different than a traditional sales guy in the Oracle sense. Some companies brand them as Customer Success Advocates or "coaches" or something similar. You aren't asking for "the sale" per se -- the customer is generally pretty onboard that they want it and will often have given their credit card already. You're just making sure that they actually get to using it successfully. This involves handholding, counseling, motivating, and being organized and persistent rather than bashing down objections, consultative sales, and always-be-closing.

In addition to it not reading like "traditional sales" from either side of the telephone, the background/inclinations/etc of the employees who do this are quite different. One obvious example: at many companies, these folks tend to run towards ladies, much like e.g. marketing teams or customer support and in contrast to how engineering teams or sales teams in B2B companies rarely tend to have large concentrations of ladies. (Positive claim not a normative claim, guys: regardless of the optimal state of the world, it is an objectively true statement that the average number of Y chromosomes per team member varies widely across departments in B2B software companies.)

As to avoiding it corrupting the flavor of the company: I'm pretty sure 37signals has a corporate culture which isn't going to turn into Oracle anytime soon, and it's unlikely that this would radically change the character of my business either. (Money generally doesn't affect people nearly as much as people expect it to.)

jasonfried3• 13 years ago

As I mentioned in my post, high-touch, full-service doesn't have to be the permanent model. Our plan isn't to do it this way forever. Our plan is to do it this way for a while so we can learn everything we can. Then later on we can automate this and make it self-service. The product will be a lot better off because of it. This was pg's point in his original post as well.

AVTizzle• 13 years ago

Towards the end of this post I realized that this was Patio11 responding to Jason Fried responding to Paul Graham...

Basically, porn.

ericabiz2• 13 years ago

Hey Patrick--as the person who (may have) introduced you to this, I should share what we're doing. We are just about to hire another full-time person to help us with this as we scale out our consulting-as-onboarding process. I promise to write a blog post once I have more data on how well this works, but I will say this: You can get a well-qualified person in the U.S. who's smart, technically savvy (not a programmer but knows his/her way around computers), and has a bachelor's degree in a related field for $40K/year in salary--and that's a step up from where many of these folks are today.

I'm really excited to have a FT person working on this, and I think in the future we'll see the "hacker/hustler/designer" trio that Dave McClure et.al. are so fond of speaking of expand into four people, with the fourth being a customer service/customer onboarding role. In fact, this is our fourth FT person on the team (other 3 are two savvy developers plus me, the business brain.) A designer will be our fifth hire.

More data: We plan to implement consulting-as-onboarding for any plan priced over $499/mo (which is our sweet spot right now), and, in fact, "do it for them" in terms of getting all their data into our system (which is where many of our customers are struggling now.) If it works and pays off in terms of customer acquisition and retention, I have full plans to do this for EVERY customer. I learned how to do this in the hosting industry--Rackspace was my example 10 years ago, and I did this with massive success with my hosting company as well. It was also one of my top tips for WP Engine when I consulted there.

P.S. Patrick, if you're still reading, we talked about doing a podcast, and this is what we should talk about. I'm game--I'll shoot you an email.

HNJohnC• 13 years ago

Is this a new writing style I'm unaware of or just plain laziness?

Get to the point blog writers!

Make your assertion clearly and right at the top, explain it if it could be misconstrued so that we know what the hell is being discussed.

Then go on to write about it.

It seems like every second blog post linked to here is a long rambling missive that only reveals the point somewhere minutes into reading it long after my eyes have glazed over.

steveklabnik• 13 years ago

... reading blog posts doesn't scale? ;)

coldtea• 13 years ago

Is this a new writing style I'm unaware of or just plain laziness?

No, seems just regular plain laziness. On your part, that is, wanting the gist of the article to be spoon-fed to you.

jonpaul• 13 years ago

I read the book Practical Typography yesterday and this quote stands out: "Attention is the reader’s gift to you." http://practicaltypography.com/why-does-typography-matter.ht...

professorplumb• 13 years ago

This is true - yet I'm pleased to see that the 37signals blog posts which are starting make it to the top of HN are those with more depth. For a while, it seemed like all of the ones making the front page were simple platitudes with no concrete examples or solutions e.g. http://37signals.com/svn/posts/3513-when-empathy-becomes-ins... .

SkyMarshal4• 13 years ago

Yes, assume your reader has just woken up and has 20+ tabs open they're trying to speed-read through before they finish their bagel and coffee and rush off to work. They want the firehose, not Dostoyevsky.

Good rule of thumb for this kind of audience - invert your essay structure. Put the concluding paragraph first, then work backwards through the supporting paragraphs till you get to the original question.

That style is much more engaging than traditional, overused, boring form of "Intro para -> 3x supporting para -> conclusion". Instead make a strong/counterintuitive/nonobvious/etc unsupported assertion right off the bat, then work backwards to support it.

That initial assertion captures peoples attention much more effectively than a slow build up, b/c they then want to read on to find out if the author can actually support it, and are gratified when you do.

rimantas• 13 years ago

They sell books. Maybe they don't want this kind of tldr audience.

thirdman• 13 years ago

I agree with you, except that beginning a piece with an argument and then following that argument with supporting evidence isn't inverted -- it's Essay Writing 101.

dragonwriter2• 13 years ago

Good rule of thumb for this kind of audience - invert your essay structure. Put the concluding paragraph first, then work backwards through the supporting paragraphs till you get to the original question.
That style is much more engaging than traditional, overused, boring form of "Intro para -> 3x supporting para -> conclusion".

No, its not "more engaging" than the traditional 5 paragraph essay, because in the traditional 5-paragraph essay format, the first sentence (the thesis statement) is the unsupported (to that point!) position which the rest of the essay supports. You aren't arguing for "inverting" the "traditional, overused, boring" 5 paragraph essay format, you are arguing for using the "traditional, overused, boring" 5 paragraph essay format in exactly the manner it is traditionally taught.

kvcrawford• 13 years ago

Interesting...there was an article on HN recently that posited a major reason for the 'failure' of the Facebook Platform was an over-engineered attempt to mitigate spam and promote quality control through automation...instead of curating apps by hand, a la the iOS App Store.

(The article in question: http://pandodaily.com/2013/07/23/move-fast-break-things-the-...)

amac• 13 years ago

human-assisted onboarding

Otherwise known as professional services. I think that's what some folks miss when they think about SAAS - it's going to be about service more than anything else despite the network delivery.

jmduke• 13 years ago

I've got two nitpicks.

1. Know Your Company is, at minimum, a $2500 service. The amount of time you can spend with a given customer is huge, in that respect, because the customer support you're doing for that one customer is proportionally much more valuable. If I have a SaaS offering billing out at $20/mo., how much time can I feasibly spend with a customer (of which there are multitudes) before I have to think of ways to avoid timesinks?

2. The "manually adding people to a database" thing seems contrived. It's valuable because you learn the departments and the employees: so what's wrong with automating the excel/email uploads and having the script spit out an org chart and dossier to the printer? (That isn't a rhetorical question -- wouldn't a birds-eye view of an organization be more helpful?)

(Overall, I absolutely agree with Jason's point! But I think the argument is a bit more nuanced than the way he presents it.)

jasonfried3• 13 years ago

1. This is part of the point. Maybe $20/month isn't the right model. Maybe there's a different model you can consider where you'd have fewer customers, but those customers would be at a much higher price point. Who knows. Point is, it's your call and it's all possible depending on how you approach it.

2. Importing the spreadsheet of customer names will likely be one of the first things we automate. It's less insightful than the other things on the list, but I still believe doing it manually is teaching us something right now.

larrys• 13 years ago

"Maybe $20/month isn't the right model. Maybe there's a different model you can consider where you'd have fewer customers, but those customers would be at a much higher price point."

Another model to consider would be to have different private label companies (that you control) serving up essentially the same product, priced differently, and offered through a different company name.

For example you have basecamp where you can manage unlimited projects at $3000 per year.

But it's possible that Boeing might pay $50,000 per year if they got even a higher level of service and a personal account manager. [1]

Now you could do this by adding a tier to basecamp but you could also be your own competition with basecamp just for the bespoke offering.

Figures arbitrary to make a point.

[1] Sold by a professional sales force or some other higher cost method than adwords.

For example I'm helping to investigate, as a favor, solutions for Physician scheduling. I'm amazed at how cheap the service is. I know that the group that wants this would easily pay much more to solve the problem in a turnkey fashion. But the current offerings are being sold on price only as a way to get attention. (And not doing a good job at it for that matter.)

alexatkeplar• 13 years ago

pg's original article was great, and it's nice that 37signals are along for the ride - but let's be clear, Jason's article is about "artisanal high-touch", not the real deal.

The real-deal looks like email lists, cron jobs, Google Docs, throwaway Ruby scripts, Access databases. It's operational and technical debt that's the least worst option. Artisanal high-touch has a bevelled drop-shadowed button saying "They paid us! Enter the transaction manually..." and full Campfire integration.

wizzard• 13 years ago

Well, pg gave several examples of "things that don't scale" in his article, and I believe Jason is referring to focusing on individual users, not doing tasks manually. He mentions eschewing automation but he's talking about automation of client interations, not automation of all the software processes.

In terms of software I agree though, clicking pretty boxes is just a different type of automation. I am co-opting the phrase "artisanal high-touch."

spitfire• 13 years ago

Nice long form sales letter on the KYC page. David Ogilvy would be proud!

krmmalik• 13 years ago

I know the article is referring to b2b SaaS products, but we employed a version of this idea on Http://storifi.it

We already knew that people wouldn't know what the game was about, so we sat with each user individually and talked them through the onboard process. We learned alot of things we would not have otherwise spotted and then capitalised on that learning. The web app was just an mvp, but we incorporated all the learning into our mobile app which resulted in a 100% activation rate.

Next step for us is now to finish off the app and concentrate on user engagement for which we'll be following a similar philosophy.

Why we're doing things that don't scale

Related Stories