Finding Product-Market Fit for New Products, with Tido Carriero, CPO of Segment
Tido has quite a long history with Plato, as we met him during our early days. We remember him welcoming us with a big smile and we are grateful to still have him on board since then!
At the time, Tido was VP of Engineering at Segment, and already had an impressive track record:
- A BSc. in Computer Science at Harvard
- 4 years of experience as both Individual Contributor and Engineering Manager at Facebook: he was part of the teams that developed Facebook Ads and Facebook page.
- 4 years of Product Engineering at Dropbox, managing up to 170 engineers across four groups.
Today, he is the Chief Product Development Officer of Segment, taking care of both product, design and engineering!
Back in 2017, when we pitched him what we did at Plato (ie. helping engineering managers become great leaders), he reckoned that some of the best learnings come from hearing other people’s experiences.
Tido naturally enjoys helping people and giving advice. When we asked him to become one of our mentors, he related on his own career and was eager to give back to other engineers.
In the course of two years, Tido has helped many mentees, was a speaker at our Elevate Summit… and agreed to do this live video AMA which we’ll sum up below!
Two of our beloved mentors led the panel to ensure a maximal efficiency:
- Abbie Kouzmanov, Product Manager at Amplitude. Prior to that, she worked in the product growth team at Dropbox, focusing on driving revenue growth for Dropbox Business.
- Justin Dilley, Head of Product at FullStory**.** He previously led a number of product teams at The Home Depot and Amazon.
The AMA was mainly focused on one question: “How do we find product-market fit, when we already have a product-market fit?” (We’ll thus use the acronym PMF for Product-Market Fit to avoid repeating it)
Indeed, while most literature about finding PMF is focused on starting from zero, what do you do when you’re already in a successful company and need to grow the business by shifting from useful to additive?
Having worked at Facebook, Dropbox or Segment, Tido shared his insights on how to find PMF for new products and features.
While he believes there’s still a lot to figure out, Tido had the opportunity to ship two new products at Segment: Personas and Protocols.
Protocols is a tool that helps companies identify tracking issues and clean up data to ensure executives make informed decisions. It helps align the whole company, diagnose data quality issues and lock specs to ensure data stays clean.
On the other hand, Personas is a tool that helps companies make sense of their customer data to deliver personal experiences throughout the customer journey. It helps increase both the relevance of the experience and revenue.
Although the products were launched in May and September 2018, they already account for a third of Segment’s revenue, split evenly between both products. That’s what we call…product-market fit!
The question is: which process did Tido and his team use to find the ideas and implement them to success?
Finding, evaluating and prioritizing ideas
Whatever your company is, there’s a chance that your backlog of ideas is as long as your arm. Shortage of great ideas is mostly a myth, while filtering them can be an actual pain. Here’s how Tido proceeds:
Great ideas are everywhere. I try to do as little filtering myself as possible. I don’t view that as my role. I think that is the role of a great PM_._
We have a product review process where if a PM gets really excited about an idea, they can put together a product requirements doc and then we have some structure in there to encourage thinking about what about the market opportunity is interesting? What are qualitative insights we have from customers? We have a structure that forces a product manager through this sort of Product Pitch. You can think of it as a structured way to discover what that opportunity looks like.
Once we do that it’s usually pretty clear to the PM which idea has the most potential. I find that most of the actual product reviews are more like rubber stamp meetings. All of the work to get there and present the thing was the hard part. Then the decision itself ends up being much easier.
Finding the right balance in terms of team investment
There are also several types of ideas, each answering a different problematic for the company. For instance, Protocols is a vision-led product while Personas is more customer-driven. How do you balance the priority in terms of investment in your product team’s resources?
Tido shared with us how he proceeded for both products he launched at Segment:
I’ll give you some background on how much investment we were making in each of them. I think Personas was a solid six or 12 months of customer discovery. And that was a much heavier kind of infrastructure investment as well… Personas is a whole piece of infrastructure that has an identity resolution component that stitches together identities from numerous different sources. And that had a bunch of fairly advanced computation that you can do it on top of that state.
Our resource investment was about four engineers for about a year, out of a team of about 50 total engineers, so call it 10 percent. A designer and a couple of product managers were involved in the early days. All in all, a pretty heavy investment.
On Protocols, the resource investment was similar. The company was a little bit bigger. It was three or four people led by a product designer and a PM in the early days. Engineers came in a little bit later there. Roughly 10 percent as well for this new bet.
Tido however believes he doesn’t have a perfect answer on portfolio allocation, as it depends on the company, the available resources and many other parameters. But this is how he usually allocates resources:
Part of the process is seeing how opportunities develop. I like to have a couple of bets going on each year. But you still want to focus at least 70 percent in your core business. I think that’s the thing everyone forgets: the core business is really important. It’s really tempting to have 10 different bets going on each of 10 percent that are building new things, but, I really like to have two or maybe three specific bets each year.
I like to test one idea that is vision-led because I think that if you want to get somewhere really different in two or three years, you have to have this bigger vision. The win for the customer problem lead is that it will be a little bit more incremental in nature and we found that it’s more of a go to market and the customers are ready for it.
The vision-led ideas take like a little more effort, as you need to show the customer the future. It’s more challenging from a go to market perspective. For those that really get Segment, the kinds of things they are able to build on Personas is truly inspiring.
To sum it up, here’s the guideline Tido usually follows:
- Keep 70% of the resources for the core business
- Make two or three bets a year on new products or services
- Allocate a bit more to vision-led products as they have a bigger chance to fail: you need to sell them to the customer, while customer-focused ideas are more incremental.
Building an ideation culture
As good ideas can come from anywhere, it’s important that a company builds a culture where people from all types of roles feel comfortable bringing them to the table. How does one ensure everyone participates in the ideation process?
Tido, who has both engineering and product experience, understands the underlying issues in the ideation process in both situations. He shared with us two great examples of how a designer and an engineer had ideas for features that were implemented into Protocols.
If you can get a designer in the mix early to conduct unbiased experiments_, that can be really cool. In the very early days we did a lot of early customer discovery. We build a janky prototype to validate our idea. We made something in google sheets and personalized it with a particular customers data._
This customer got really excited and told us this is exactly what they needed. We showed it to a couple more customers and that very early stage validation can turn into an actual product.
This is a good example of someone who is awesome at customer research, thinking about the cheapest way to show someone something and see if they like it. It’s interesting to me how similar the janky prototype and the actual product can turn out if you do the customer research right.
The other story is about our feature called Typewriter_. In Segment, you enter your schema of the data you want to collect. We heard from one of our customers that the problem is when you are tracking the event, you want to make sure the event gets implemented correctly at the code level._
We didn’t want to build a code editor for this. One of our engineers worked with the customer on this and was able to hack out Typewriter. It basically plugs into a code editor. When someone is working on implementing Segment tracking, they access specs that respect what they have specified into Segment. If you make a mistake it will red underline and it won’t compile under certain rules.
I don’t think we would have made that connection had an engineer not been there in the room. Only the engineer knew this and he had to be exposed to the customer to make the connection.
The best way to ensure everyone in a company find relevant ideas is to let expose them to the customer directly. Be it engineers, product-people, sales execs or marketers, engaging with the customer is plays an essential role in getting to PMF.
When to spend time with customers
We’ve seen that PMF happens when employees talk to customers. When, and how much time should engineers and product managers spend time with customers?
I think the answer should be “a lot”. There are different phases of a product, 0–1 and 1-n. During 0–1 this should be all the time. What is the fastest way you can get to the next learning. Sometimes you need to write code to do this but you should make it quick. You should be talking to many customers a day.
When you get into the 1-n phase, it depends more on the product. If it’s a backend product to improve speed of something, you probably don’t need to check in as much because there won’t be many relevant new learnings. Something more customer engagement may require more customer contact through this stage_._
Detecting Product-Market Fit (or moving on)
How do you know when you attained PMF, and decide that what you shipped is good enough for you to move onto the next thing?
Before shipping Protocols, Tido and his team actually worked on two other prototypes, both of which were unsuccessful, yet an important part in building the right solution. A little history:
We knew data quality was the important issue on people’s mind. We had three fairly different prototypes and two fast failed. For the first one: we had this theory that anomaly detection was the biggest thing. Ultimately, we felt a lot less pull with anomalies and had it set up for 10–20 customers. We realized there are other companies that do this much better than us.
The second one was transformation : taking the event name and migrating it to a different event name. It’s actually very hard to do this migration. We thus decided the requirements were a much bigger scope than we anticipated. We determined transformations would be valuable but needed to be pushed out 6 months.
Basically, the two other prototypes were worth the time. There was one memorable experience onsite at Fox and our PM and designer shadowed the QA engineers to see how they do their QA. This helped our employees to know what we could help them with. So it was a combination of this onsite and seeing the opportunity and getting feedback on the prototype.
In order to consider an experiment as a failure or a success, it is crucial to have some success criteria in mind:
You need to establish success criteria up front. With Personas and Protocols, the criteria we had set up when we started it was to increase average contract value. That was thematically what we were interested in doing as a business last year. Attach rate and attach amount.
Some of the other Protocols features that we considered adding also just felt like things that we wanted to include as part of the core platform. It was part of that North Star metric that guided us in choosing between the three prototypes. We do have a lot of other stuff going on where the North Star metric is not at all governed by revenue.
Building on product-market fit
Once your new product or service has attained PMF, it’s crucial to consolidate, improve, and build on it. Here is how Tido and his teams prepare for that step:
The ethos of team is that they prepare for the next step up. We are hoping to have engineers work full time on developing new prototypes associated with Personas. Those are projects that are important otherwise you will have an incredibly leaky bucket. You need to continue to scale for bigger customers, etc, but the more you can keep the culture of the MVP and the culture of the 0 to 1, the better off you will be.
The best teams are always doing MVPs and that’s how we think about feature development.
Here are the highlights! In a nutshell:
- Build a culture where people aren’t afraid of sharing ideas
- Make sure employees spend time with customers
- Evaluate ideas and prioritize accordingly
- Define a North Star metric / success criteria
- Allocate the right amount of resources to build an MVP
- Try again, fail again, fail better
- Validate success
- Consolidate, improve, build on your success
We hope you enjoyed these AMA highlights on how to get to product market fit in an already successful company, and would like to thank again Tido, Abbie and Justin, as well as our community.
We stole that one to Samuel Beckett, but it fitted perfectly here.
We don't have any blog posts yet.
We are doing our best to find what you are looking for. Don't hesitate to contact us if you can't find what you need.
You're a great engineer.
Become a great engineering leader.
Plato (platohq.com) is the world's biggest mentorship platform for engineering managers & product managers. We've curated a community of mentors who are the tech industry's best engineering & product leaders from companies like Facebook, Lyft, Slack, Airbnb, Gusto, and more.