A few weeks ago, Paul Graham wrote a great essay about doing things that don’t scale. As usual, it’s clearly argued, well written, and full of solid advice. If your first instinct as a product developer or entrepreneur is to automate everything you can, I recommend reading Paul’s piece.
I wanted to add something to his article. While he’s mostly talking about doing things that don’t scale when you launch a new business, I think this advice is equally valuable when you’re launching a new product in a well-established company. In some ways it’s even more valuable in that situation.
There’s no question that automation is enticing – especially for software companies. No one disputes it’s easier to scale your business when you have machines doing the work.
But automation can also lead to myopia. And premature-automation can lead to blindness. When you take human interaction out of a system, you’re removing key opportunities to see what really happens along the way. You miss stories, experiences, and struggles – and that’s often where the real insights are hiding.
A real-life example at 37signals
We recently launched our new product Know Your Company. From the start, we decided to approach development, design, marketing, and sales differently than we have in the past with our other products. We wanted to start out by doing lots of things that we knew wouldn’t scale.
Prior to Know Your Company, every product we’ve ever designed has been self-service. People found the product themselves, signed up themselves, paid themselves, and got started themselves. No human interaction required. It’s obviously worked well seeing that this is our 14th year in business.
But all this automation has also created some institutional laziness. We have systems that work, so we just keep reusing them when we introduce new products. Why think about how we bill people if we can just plug in our own centralized billing system that we already have? Why re-think how we administer customer accounts when we already have a centralized admin screen that runs our other products? Why re-think pricing when already have a monthly subscription model that works?
There’s something about all that comfort that makes me uncomfortable. If you haven’t reconsidered the fundamentals of how you do business in a while, you’re probably past due. And hopefully your late fee isn’t going to put you out of business.
So with Know Your Company we wanted to reset our assumptions and eliminate a lot of the automation we usually lean on. Rather than separate ourselves from the customer, we wanted to bump into the customer as often as we possibly could.
Specific examples of things we’re doing manually that could have easily been automated
1. Sales and demos. If you want to buy Know Your Company, you have to be willing to hear our 30-minute pitch and follow along as we give you a personal product demo. There is no sign up page. There are no product shots. There’s no fancy video tour. There’s just a link to my personal email account. Drop me a note, let me know you’re interested, we’ll trade an email or two, and if we think you’re a good fit for the product, and the product is a good fit for you, we’ll schedule a demo, talk, share some stories, and show you the product.
WHY HAS THIS BEEN VALUABLE TO DO MANUALLY? There’s nothing quite like seeing or hearing the person you’re pitching. It’s never just a demo – it always leads to a little conversation. You get direct questions and you can give direct answers. You can tell a story, listen to their stories, share personal experiences, show specific examples of how you use the product yourself, etc. Emphasis can be placed on different parts of the demo based on how they’ve reacted to something earlier in the demo. Further, in just 20 minutes, they get to see every single nook and cranny of the whole product. No stone is left unturned. You can make sure they know it all right from the start.
2. Adding employees to the system one at a time. To use Know Your Company with your company, you need to add your employees to the system. Instead of making you do that, we do it all for you. You just send over a simple spreadsheet with your employees’ names, email, title, and department, and we’ll add everyone manually to the system for you. And if you hire someone new, and you want to add them to the system later, it’s the same process – just shoot over their name, email, title, and department and we’ll add them for you.
WHY HAS THIS BEEN VALUABLE TO DO MANUALLY? Being forced to enter each person manually helps you understand the types of people who will be using the system. You know (and type) their names, their titles, their emails, and their departments. There’s not just 32 people in this company, there’s Jim, Lisa, Aaron, Bryan, Jennifer, Mark… There’s a few people in customer service, a few account managers, four graphic designer, etc… It also reminds you that data entry can be error-prone and a real pain in the ass. That’s a good thing to remember when you’re usually in the business of asking people to fill out web forms with lots of fields.
3. Four stages before the product starts working. When someone signs up for a product, it usually starts working immediately. That’s because it’s automated and automated things are usually optimized for speed. But we wanted to slow that process down to four specific steps, and display those steps along side customer accounts in our admin tool. This way we could see where a specific customer was in the process. In order for someone to go from one step to the next step, we have to click/toggle the flag in our admin.
WHY HAS THIS BEEN VALUABLE TO DO MANUALLY? Having four distinct stages forces us to maintain contact with the customer every step of the way. This allows more time together, more opportunities to build trust, more chances to show them that we’re reliable and always around to help. This handholding may seem excessive, but it’s really not. All four steps usually get done within a few days. There’s not much to it, but even that little bit goes a long way to make the customer feel more comfortable with us and with the product.
4. Having the customer send out an announcement to their company. Part of the sign-up process requires the customer to make an announcement to their whole company letting everyone know why they’re using Know Your Company, what’s in it for the employees, what’s in it for the owner, the goals, the intended outcome, and how the system works so no one is surprised when they interact with it for the first time. We’re happy to coach them through the announcement, or even help them write it if they request, but we won’t turn the system on until the customer has explicitly told us they’ve made the announcement.
WHY HAS THIS BEEN VALUABLE TO DO MANUALLY? It would certainly be easier if the system sent the announcement automatically, or if the customer just had to click a button to blast out a pre-written email, but it wouldn’t be anywhere near as meaningful. By asking the customer to make the announcement however they usually make company-wide announcements, their employees will pay more attention. Some owners make company-wide announcements in person at an all-hands meeting, other post announcements to Basecamp (or something similar), and others do it via email. The point is that when they do it their own way, instead of our automated way, it has more weight and carries more meaning.
5. Generating invoices and marking them paid. Whenever there’s an invoice to be paid, we have to manually click a link to generate the invoice. Then it’s sent out to the customer via email. When an invoice is paid, we have to mark it paid manually, too. So every time we want to bill someone, and every time we get paid, we’re directly involved in the process.
WHY HAS THIS BEEN VALUABLE TO DO MANUALLY? I think it’s important early on for everyone who’s working on a product to be witness to the financial success (or not) of the product. Further, it just feels good to click a link send an invoice and later on press a button to mark something paid. Certainly automated reports could fill you in on the raw numbers, but it’s really a whole different thing to “work the register” yourself. We also pipe a “XYZ Company just paid an invoice for $2800” into our Campfire room so everyone on the project knows we just got paid.
What’s a little inefficiency worth to you?
So remember, efficiency shouldn’t always be the goal. Especially early on. If you want to learn, you have to struggle. Automating away all the struggle may teach you how to be efficient, but it won’t teach you much about your customers.
And of course at a certain point it does become worth the trade. When it’s time to be efficient, automation does make sense. I can see us automating a lot of the manual parts of Know Your Company later this year. But right now? Not yet – we have a lot more to learn.
Lars Klevan
on 25 Jul 13Nice post! Minor typo: efficiency shouldn’t always be the goal.
-JF: Fixed. Thanks.
Greg G
on 25 Jul 13Not to be a hater, but I still don’t understand what Know Your Company is.
I want to be able to watch a 30 SECOND demo first, so that I can get a sense of what Know Your Company is, before I even consider exhausting the time to even setup a demo, let alone dedicated 30 minutes of my time to be pitched the product.
Said another way, the leap between me not understanding what Know Your Company is because I can’t even see it offline, on my own time … to jumping into the hassle of setting up a 30 minute sales pitch just isn’t inviting enough. I’m busy. I have 41 employees. And I consider my time a limited resource, so I don’t set aside time during my workday unless I have feel it of value.
Please create a 30 second demo.
Jason Fried
on 25 Jul 13Greg, you aren’t being a hater. You make a fair point.
But this is all part of the pitch. I’m looking for highly motivated customers who read https://knowyourcompany.com and just have to get in touch. So far nearly 100 have, and and nearly half of them have purchased the product. We’re super happy with those results so far. If we can continue that we’ll be in great shape. Filtering and self-selection through a moderately high bar helps make sure the people we talk to are very serious customers. This insures their time, and our time, is well spent.
I know this strategy might leave out some people who potentially could be customers but want to know more ahead of time, but that’s fine with me. Especially early on. We’re not looking for every customer, we’re looking for customers who absolutely know they need something like this and they are willing to invest 30 minutes of their time to get a personal in-depth tour.
Down the road, if/when we automate these things, and post a tour, and maybe share some videos, then those who couldn’t spare 30 minutes will be able to watch and make up their own mind on their own time.
Karthi
on 25 Jul 13Good insight! Let KYC grow. Cheers!
Greg G
on 25 Jul 13@Jason Fried
I understand your new approach to this service and I respect that.
Just keep in mind, there are people like me who simply see something and want to instantly ‘buy’ it.
And if I feel like I have to be “pitched” something, regardless of what that is (car, home, etc), I will simply avoid it – even it it’s a great product/services.
I like to make decisions on my own, and on my own time … without dealing with someone in “sales”.
So when I recently bought my new Audi S4, before I selected the product, I spent considerable time online researching different makes and model. I was never pitched the car, because I evaluated it on my own, on my own time.
So your new approach, which is commendable, might be alienating a huge population of potential customers who simply don’t want to be pitched (nor maybe don’t even need to be pitched because they are already sold on the product).
Take this as an example, I put a lot of value in what my core group of close friends recommend. Some of my friends are business owners themselves. If one of my friends used Know Your Company and said “Greg, you have to use this tool. It’s great”. I simply want to go out and buy the tool. Not be sold on it by 37signals through a 30 minute pitch … because I’ve already been sold on it by my friend.
So you’re going to lose the ‘network effect of recommendations’ by using this sales approach.
Again, not to be a hater. I’m just trying to be a friend and point out potential negative aspects of this approach from my point of view.
(Plus, the one time payment it’s really a huge selling point to me vs monthly subscription. Frankly, it makes me uneasy because I think to myself – how in the word are they going to stay profitable with that business model. If they terminate this product, I just lost all of that time I’ve spent using this product)
Jason Fried
on 25 Jul 13Greg, I understand everything you’re saying. We sell all our other products in the exact manner you describe. I have a lot of experience making things that sell themselves. You don’t need to tell me how well that works ;)
But right now we aren’t selling Know Your Company that way. Some of the reasons are described above in the post. There are other reasons, too. And as I mentioned in the post, we may make it highly automated and self-service down the road, but for now, we’re perfectly comfortable taking our time, one highly-motivated customer at a time.
Greg G
on 25 Jul 13@Jason, that is not correct.
Breeze was sold in the manner similar to Know Your Company (one time payment) and was shutdown after just 4 months.
DaddyGeek
on 25 Jul 13Have to say I agree with Greg too. I also hate (with a passion) the whole sales letter approach, it’s cheap – and to be honest I was expecting you to try to sell me an ebook or something.
I wouldn’t be surprised if that site was doing 37S and your other products more harm than good – it’s something I expect to see from a chancer, or a company in financial trouble.
BS
on 25 Jul 13I love this concept. Sadly, my company is not large enough (yet) for “Know your company”. I will be reaching out to Jason when it is. However, we’re not going to rush growing for the sake of growth, so it may be awhile before we are able to see a demo.
Jason Fried
on 25 Jul 13Greg, to your point re: Breeze… I was commenting on self-service, do it yourself signup. That was the crux of your argument and the point of my post.
Last thing I’ll say about that is… Breeze was around for a few months and only generated about $10,000 in sales. Know Your Company has been a product for about a month and we’re fast approaching $200,000 in sales. So very different scales. Also, the “recurring” part of the revenue model is that companies have churn and grow. So as they add people, they’ll pay $100 per additional person.
Greg G
on 25 Jul 13@Jason
Are you stating that if a company has an employee leave, then they hire a person to back fill that position – the company will have to pay another $100 to 37signals … even though their total employee count has stayed unchanged?
Jason Fried
on 25 Jul 13Greg: That’s correct. It’s $100 per physical person, not per virtual seat. We’re not selling slots or spots that can be inhabited by dozens of people over time. This is about investing $100 in a specific person to get very specific feedback from them for as long as they work for you.
Lisa Simmons
on 25 Jul 13Great post! Helpful and awesome information. Thanks!
LS – http://porsche356thefirst.org/
Greg G
on 25 Jul 13Ouch.
That sales model is unheard of in software and cost prohibitive.
And puts unnecessary burden on me as a business owner to keep track of writing you $100 checks for every employee I hire.
Steve
on 25 Jul 13Jason,
I don’t love the idea of paying per “seat”. This makes it more difficult to do a product “Sun Setting” (like you did with Basecamp classic).
What is great about Basecamp classic, is that people continue to pay money for your legacy product. This supports the server costs and allows you to keep the service up for a long, long time.
What happens in 8 years when you want to “sunset” KnowYourCompany? You don’t want to keep adding additional users and you have a crap load of old users that aren’t paying you any new money.
It would make it much harder for me to want to keep up a legacy product when I am not making any new money on it (but still have to keep the servers, customer support, etc..) going.
Thoughts on this?
David Andersen
on 25 Jul 13Greg, just to jump in here – I don’t think you should be framing this as a software purchase. It’s a service, that happens to be delivered via software. And I’m sure you realize that every time you replace an employee with another you have acquisition costs for that new employee. Why would this be any different?
Also, with respect to sales people and getting pitched, I’m with you. Far too many salespeople are bad at their job and don’t really listen which means they are 100% focused on pushing product. But occasionally I encounter a sales rep (like my VW salesman) who could be writing for Car & Driver, such is his depth and passion. Despite my research I was still learning things from him the day I drove the car off. So while I avoid sales people as much as possible I’d hate to miss out on that sort of interaction, rare as it seems to be.
Joshua Dance
on 25 Jul 13I think this is crazy, and brilliant. Forcing yourselves to see and handle every step helps you learn about your customers. I read you article about the grocery store owner you know and I was wondering how you would apply that on the web. Do it manually until you know enough to automate is a great strategy for certain situations. Good luck, and thanks for sharing.
@Greg G
on 25 Jul 13So let me get this straight. You raised a concern with Jason about the “pay once” model over an on-going subscription. You said you would not want to pay once for a product that may not be around due to lack of profitability. Jason, comes back and tells you that there is re-occurring revenue generated when businesses grow by them adding new people (not seats).
Jason addressed your concern and you still come back with “Ouch!”. You started out stating “Don’t want to hate” and “as a friend”, when indeed you intend just the opposite. Originally you seemed a sensible person asking good questions. Now, you have exposed yourself for what you are, a troll who should be sitting in a corner with a clown hat on.
Amrita
on 25 Jul 13I think your approach is right on. I work with a lot of early stage startups who want to deliver their product in a low-touch, SaaS model. What ends up happening when the product is first released: their prospects get “stuck” along the way somewhere in the trial process, conversion numbers are way below target and the company tries different experiments in messaging, pricing, you name it.
Every time, I suggest they try talking to some of these folks and holding their hand through the onboarding process. And every time this leads to a number of lightbulb moments around why people came to try the product, what their expectations were, what they think this product can do, what made them switch from what they were doing before to trying this (h/t #JTBD), what their experience was like, and where the gaps were that caused them to either not use the product or not convert to becoming a paying customer.
I think people are missing the point that a manual process is really helpful when trying to get your first 50-100 customers. Not necessarily the next 1,000.
I’d argue it’s actually less efficient and more expensive in the long term to start automating everything out of the gate.
Devan
on 25 Jul 13Interesting – I’ve spent the past 20 years selling Small Business accounting & ERP systems, and this is pretty much the exact same sales model we use. Distinctly old school.
I’d be interested to hear more feedback in 12 months or so about the ‘old way’ vs the ‘new instant self service way’ to see which one is working. It will be quite ironic if we see a trend back to the old way of selling ‘solutions’ that was espoused by the ‘consultants’ at IBM and Digital etc. back in the day.
Peter
on 26 Jul 13If you have $200,000 in sales with 100$ per employee and avg. 40 employees per company then KYC exists for 50 weeks already?
Des Traynor
on 26 Jul 13Really interesting stuff.
Two things jump out at me:
1. Requiring the announcement is a pretty genius move.
Every tool that does a job in this area (HR, comms, activity feeds, even project management) struggles from company adoption (aka engagement). By mandating a broadcast announcement you’re setting the product up to win.
2. Your approach forces you to revisit old assumptions
This is both refreshing and revitalising. Using new technologies, new billing, new sales techniques, new support, etc. teaches so much.
I’d be interested to know if any of these learnings might find their way back into Basecamp. e.g. For a $20k/year deal would you go down this road of direct sales, etc.
Carsten
on 26 Jul 13I’ve just read your interesting approaches on knowyourcompany.com. It’s really funny that the problems you were facing when your company grew are almost identical to the ones we had in our daily agency life – although we’re just a little rat-shop. So I think the core problem, the lack of (valuable) communication and recognition, is not only present in growing companies, but also in small teams.
For this reason, we had built a simple tool (“productivity and motivation log book for teams”) for internal use last year. After some months of consequently writing brief “day reviews” by all team members, the motivation started to increase noticeable. In other words: The team spirit came back. (Therefore we decided to release our tool as “teamspir.it” recently.)
Of course, the “regular team writing approach” is much simpler than the story behind KnowYourCompany, but it can be a good start for small teams (up to 10 members).
Jim
on 27 Jul 13Excellent post Jason. So much to learn from talking to “live people” as the product is developed.
To Greg: you make some good points- but after all this wouldn’t it have been best to have got a one-on- one product demo? I’m sure hearing from Jason why each new person has a cost attached would have made perfect sense.
Michael
on 27 Jul 13There is so much to like about the pricing model besides that it scares away tire-kickers.
Traditional SaaS is a consistent monthly fee with bottlenecks to force upgrades as a company grows and only a loose incentive to grow the business of the customer. Revenue-wise, front-end costs are recouped over time and many customers tire-kick and don’t stick around for long. Using the traditional model, KYC would probably cost $50-$150 per month. A customer might upgrade a plan after a little while, or they might quit after a month.
KYC has no bottlenecks; customers get all the features for their employees. The money paid to 37S upfront represents perhaps 2 years of monthly payments using the conventional model. But, and I believe this is huge, KYC is a product designed to help a company grow in size, and the cost of KYC is directly tied to growth. KYC isn’t a bet that a 25-person company will hire 3 people and replace 2 in the next couple years for $500. It’s a bet that a 25-person company will unlock a great deal of potential in its employees and use that to become a 50, 75, 100 person company (for $2,500, $5,000, or $7,500 more) before growing out of the product as is always 37signals’ preferred endgame. KYC has the potential for more recurring revenue than traditional SaaS in addition to an initial payment, and it’s done in a way that feels better and more fair.
KYC is an optimistic product directly connected financially to the customer’s success in a way that also lets the customer perceive and even appreciate that direct connection. I am so excited for this model to work and spread to other software companies.
Nate Tuganov
on 27 Jul 13Jason, I’m too agree with Greg. I can’t afford to spent 30 mins just to listen or get the idea of what you are going to sell to me. Show me the case first, a real one, then, and only then, if I’m interested to and will ask for demo, we can speak.
I agree and understand why you want to pitch, mostly, users don’t read, don’t watch much and their experience with the service is very low, they churn because they don’t get the service, don’t get how to optimise their business process. However selling a pandora box and spent my, our someone else’s time is a wrong idea.
I really do believe that showing even a simple case will increase leads and thus sales.
Regards, Nate
Jason Fried
on 27 Jul 13Nate, as I’ve explained, this isn’t about more leads and more sales. It’s about the right leads and the right sales. Quality, not quantity.
We’re a small team of two people giving all the demos and pitches right now. A dozen more leads a day would be a bad thing, not a better thing.
Could we hire more people to give more demos? Of course. But that’s not what we want to do right now. We want to pace ourselves and keep it manageable for the team we have.
The filters we have in place (you have to be up for a 30 minute demo, ~$2500 minimum cost, etc) helps reduce casual tire-kickers and leaves only highly motivated, qualified leads.
Remember, we know how to move fast – Basecamp sees around 6000 new signups every week. If/when we’re ready to put on the afterburners with Know Your Company, we can do that.
Until then, we’ll be working closely with a small set of highly motivated customers to insure they get the most out of Know Your Company.
steve moulton
on 30 Jul 13Interesting thoughts. I agree that you can over-automate and lose touch with your customers. I also agree that in the early stages having initial users educate you is key. I don’t know whether something this time intensive (on both sides) is the right approach though. I am one of three founders, two of us answered 9,500 helpdesk tickets during the first two years. I think it served as a good way to interact with our users and also learn about what our product needed without insisting on handling every sign up.
We also have an ‘add team members’ piece (csv import). We made it very clear (we think!) that if you needed help with the import, we were there to help. We also triggered an automated email to one of the founders if an import failed…(many xls import attempts!), but if you could manage on your own, by all means go ahead.
I get that this is a little different, you want to add members so that you know what role they play in the organization, although I’m not entirely sure, a name, email, photo and title will tell you that.
I suppose a large part of this comes down to the resources you have available to apply to a product. 37S is sufficiently resource rich to try this approach and see how it works. Good luck with it!
Drew Meyers
on 01 Aug 13Really brilliant strategy & pricing, if you ask me. It means that the business owner paying has a vested interest in getting that specific employee they spent $100 to setup to use the service. With saas models, churn happens when people don’t use the product for one reason or another – even though it can help them immensely. If I’m paying $100 per, then I care a lot more if my employees use it than if I’m paying $150 or $200 per month for my whole organization to use it (or whatever that # would be).
This discussion is closed.