the death of the two pizza team

why teams will get bigger, more cross functional, and employers are going to pay more for each role in our the LLM future.

tldr; in the future teams will likely be a dev, pm, designer, sales rep, success rep, marketing person and a brand person

as an engineer i can absolutely design, ship, deploy, then iterate and track a feature by myself. this leads smart folks to think “I guess teams will be smaller”. the truth is, i could always do that. i did that. every startup was exactly that at one time. we didn’t add pms, design and all the process because a small team couldn’t complete the task. we added it because the more we build the further from the customer we get. the further from the customer we get, the worse decisions we make.

the folks espousing small teams are coming from a good place. the theory being people can do more, so small groups can tackle more, so we’ll have small teams like a designer, a PM and an engineer delivering work. this makes sense initially but works from the assumption that the problem relates to how fast we can ship features. that’s never been the problem.

the problem has always been shipping the right thing, never shipping the next thing

most saas companies are a view into a database that act as a system of record. that’s it. we spend a lot of time dressing up the database so that people can interact with it. if we have some kinda moat we do some processing magic on the data and then put it in the database and then dress that up so customers can see it. there’s a persistent belief that putting the data in front of the customer in some elusive right way will make the product good. but it’s actually the system of record stuff that’s valuable (and increasingly the post processing stuff). it CRUD all the way down. (i’ll maybe write a future post about how the next thing saas will sell is religion)

now (if our new AI overlords are to be believed) one person can do so much more we’ll have a single person billion dollar business. i suppose this is true, i expect it will happen. but i don’t think it’s an important change. i have been able to single handedly ship the system of record stuff and make different views into the data for a bunch of years. great right? WRONG. the value isn’t the view into the system of record. it’s owning the system of record (and that post processing stuff).

building that next feature was never the value. it was never the problem. if you were one feature away from a successful saas you were a dead company walking.

okay i believe you owning the system of record is good SaaS, but can’t everyone build that?

yep. not everyone will and actually getting customers is super hard. but, just having a db and view layer into isn’t a differentiator, it never really was. people thought it was because saas has fat margins and fat margins make you lazy. if you want to see the future of saas you want to look at businesses that are in low margin spaces and are somehow winning in other ways. obviously the grocery driven development (see blog post i never finished) guy is going to say this but, look at grocery stores. you have a favorite. they all have aisles and produce. but you like one because of the people, or the lighting. maybe they have some more fringe products or just fewer choices.

there’s a mix of brand, design, product, price and convenience that makes people like what you offer, just like grocery stores. you need to thread that needle like wegmans or trader joes or whole foods or dollar tree. then your customers will be satisfied (whatever that means to your customer).

the thing is that one or two people can’t do that. your super powered llm pilled engineer can’t ship that mix. your uber designer product maker can’t keep it all together (especially as all saas converges on all features). you need to be mainlining what the customers are thinking, feeling and experiencing. how do you do that? you fucking talk to your customers. constantly.

that jacked zyn powered golf enthusiast over there dialing for dollars all day, they’re one of your best weapons. they’re talking to customers, selling your product all day. all those little objections, that’s what you need to differentiate on. the gen z tik tocker in customer success who knows exactly what the current meme is, they are your brand voice. they talk to your customers constant and know exactly how your customers feel about you. they know what you should build next.

we built engineering teams to look something like three devs, a designer, a pm. that made sense when releasing the next feature was slow. but now, releasing the next feature can happen quite fast. the aisle is full of choices, the differentiator is building the right thing, the right way, and communicating it through all the noise. the five person engineering makes no sense now. we need to be closer to the customer, more in tune with the brand and marketing we need to work together more.

:imo: the feature development cycle will look something like this. the pm identifies a need. brand and marketing run an instagram ad around that idea. that gets validation via attention, clicks, whatever. if it pops right, engineering builds it. data can jump in and add some tracking if it makes sense. success and sales immediately get on the phones upon deploy. does it sell, is it solving customer need, that kinda thing. said another way is this earning its shelf space? then design starts. design and product now figure out if this fits the vision, aligns with company vision, brand.

is this just letting sales dictate product and engineering?

kinda. you lied to yourself when you thought there was ever something else, it’s always been that. it’s what the product development loop can and should look like. it’s not the iterate bs that people espouse. and it’s not some goofy agile idea that sells great to execs that don’t want to take responsibility for anything. it’s actually talking to customers. constantly. collaborating with the folks that are closest to our customers and just making that, is a pretty good idea.

empathy remains an underrated skill. the best designers i’ve worked with score high on empathy. the pms who take in all the need and shape a solution, they’re doing that with empathy. well, the sales and success folks have been largely left out of this loop. for good reason. they’re busy and i was busy as an engineer too. we’re all still busy, don’t get me wrong. but that success rep who really gets our customers knows intuitively about some big wins.

let’s imagine a feature that exists today in your backlog. that the “email as pdf” button request that nobody has prioritized. it’s a paper cut. our success person knows the reddit threads, customer forum posts, facebook groups, and the emails that have asked for this for years. we can build that feature on a monday and release it on a tuesday now. we should probably bury it in the UI somewhere (design can help solve that), but we can fix that paper cut. those customers now have one less reason to churn.

if you make a product for everyone, it’s for no one

if we solve every problem, and add a config for our product will suck. we need taste. the word taste was popular a while ago but everyone decided it was too hard so now we talk about something else. we shouldn’t build some things. restraint is good.

i don’t really know where the line is. product and design alone don’t know either. marketing, sales, brand. put all that stuff together and you have a company. the feel, voice and identity of your company, that’s the thing you want to build. to do it right you need all those voices in a room so everyone’s intuition about what it is we’re building can grow.

this means that the sales person, needs to learn enough about engineering to communicate. they need to start understanding what might be hard and why. the advertising person on your team will need to have some sense of why a feature might take a long time to build (even with the robots). then they can position the ad correctly. those people will cost more. this is a good thing. setting a higher expectation bar and enabling your people do more is great, it’s good for your product, customers and bottom line.

yea but i need my sales and success people doing sales and…

no you don’t. go into a trader joes and ask some random employee where the quinoa is. they’ll freaking walk you to it. the identity of their company is we all are responsible for customer experience. in a world where we can build anything in saas we need to think in a similar way. we’re all responsible for the product, we can take an hour and have a meeting about what people are hearing on the phones. we can send a slack message. those changes will enable us to build the features that are our brand identity for our customers.

Sign up for...