How Pipedrive reached 50,000 paying customers
Earlier in the year Pipedrive crossed the 50,000 paying customer mark, an event we celebrated with
cheap champagne insightful stats. I thought it’d be a good idea to follow up my “how Pipedrive got to 10,000 customers” post. A couple of things are exactly the same, some have lost their relevancy and there are several new themes.
What follow are my observations of things that got us from 10,000 to 50,000 customers, in no particular order, but leaving the most important thing last. Note these are observations of a marketer, if you asked one of our product managers, sales leaders or investors to write the post, you might get a different post.
Where it starts and ends: Pipedrive has a great product
One thing that hasn’t changed is that in order to grow you need a great product. A great product solves a problem, is well-designed and runs without glitches. Several direct copycats and many view or feature clones later I should just add that product vision becomes increasingly important as you grow. It’s easy to copy designs of individual views, it’s much harder to copy workflows that are based on a strong understanding of the most effective way to get something done.
We provided great support with a human touch
Pipedrive’s trial-to-paid conversion increased notably after we had brought on more people to improve support response times and quality. Since then we’ve made lots of changes relating to support, altered the channel mix between phone, email and chat, used different tools and CSAT definitions, scaled from one to more than 50 support people – and all the while maintaining a human voice, which greatly has helped to simplify my job as a marketer.
Our Head of Support Martin has written more about his/our journey here.
We focused on recommendations and findability, or the Two Hedgehog growth model
Here’s a bold but not very controversial claim: the two most important ways of discovering new products and services in many categories are: A. talking to friends or colleagues and B. searching online.
I’ve learned this over and over again, whether by talking to people or asking “how did you hear about us” in signup flow.
When we at Pipedrive started testing different channels, I wasn’t fully conscious of that, but by nature of intuition or luck, a lot of our new channel experimentation and 90+% of the results have come from encouraging recommendations and increasing findability.
A product with a high NPS score is the key for recommendations. On the marketing side, we’ve been able to add some layers of growth by building a referral program and optimizing triggers that remind users to spread the word.
Us, marketers have more control over findability ie. making sure that as people do their research they find you no matter what words they use and no matter where they click on. Which is relatively straightforward. First, keyword research tells you what searches are most relevant for you. And then there is the invigorating task of getting listed on as many positions as possible for relevant searches on the first page of Google*. We needed a combination of SEM, SEO, content production, ad buys, review management and media relations – and while we have some of these pieces figures out, this job is never “done”.
More than any particular tactic, the overall focus on things that drive recommendations and findability has helped Pipedrive grow. There’s a separate longer post about this approach here.
*In the SaaS world, searching online usually means Google or another friendly neighbourhood search engine, in other categories, it may be Amazon, Google Play store, Youtube, etc.
We talked to customers and learned our “jobs”
How do you make sure that what customers want is the same that you build is the same that you sell? At Pipedrive we’ve tried different approaches for determining the epicenter of our efforts. We started with the vision of founders and senior team members, as you do. Doing a persona study was helpful, the process of going through 35 customer interviews in a short time even more so than the result. Even AB testing the message provided some value in helping to understand what we’re about. But something felt missing.
Then we stumbled on jobs-to-be-done as a framework to understand people in buying (or not buying) context. Inspired by Intercom, Matt Hodges, Anthony W Ulwick and other resources, and guided by Alan Klement we ran our jobs research and the result was like getting a really good hiking trail map. You can still get hopelessly lost, but if you do, it’s only due to your own stupidity.
Arguably, JTBD didn’t play a critical role in getting to 50,000 customers as we completed the research only towards the end of last year and still have some way to go in implementation. But it would have taken longer to get here without our previous attempts and without talking to customers.
When we had too many ideas, we added ICE
As we grew the team we increasingly got into situations where we had too many ideas to execute. And often they were not necessarily great ideas, so we wanted to learn to prioritize the different ideas for new paid campaigns, growth projects, AB tests, etc.
We tried several flavours of prioritization: dictatorship (someone making the decision), democracy (voting for the best ideas), disguised randomness (the person with the best-looking spreadsheet making the decision) and loudocracy (the person shouting loudest making the decision). We had a tolerable hit rate, but we knew we were lacking a good methodology.
At some point, I came upon I-C-E framework which is a dead easy way to weigh Impact and Effort against each other, and add the seemingly insignificant but in reality difference-making Confidence. My teammate Aivar has summarized our approach to ICE in this post And it’s worth adding that ICE is not something that is useful just once per quarter but kind of all the time.
When one of my colleagues wanted to make a new joiner understand his priorities and used the words “I’ll just ICE my to-do list for them” I knew we were getting close to using this framework enough.
We found platforms (before others did)
This one is basically a copy-paste from my earlier post, but it’s still super relevant. During 2011 Pipedrive got almost 1/3 of new signups from Google’s Chrome Webstore. Not bad for the little amount of work required for the marketplace listing back then. As most competitors were slower to appear there, we enjoyed a period of very low competition there and so Chrome Webstore made a huge difference to the growth of the company at that stage.
There are almost always new emerging platforms you can use to boost growth. In 2012-2013 it was very valuable to be listed on GetApp, for example. While the volume was low it drove highly relevant traffic at high-ROI rates.
We acknowledged that some of the most effective marketers are not marketers
And if you know you can increase signups by 50% by coding up a Chrome Webstore connection, you’ll also know that this is way more effective than most things a traditional marketing team can generate. For a long time, we had lots of ideas what else we should build that would be “engineering as marketing” but these ideas were competing for the same resources that were needed to build core features, so little got done.
Our first step was asking a few engineers to work on the website and tracking most of the time, something that has grown into a full-sized team by now. After raising our A-round and scaling up core product engineering we felt we could afford the luxury of adding another engineering team, this time dedicated to our non-web growth initiatives. In its first year of existence, this team had produced one “hit”, one dud, several medium-sized successes and a good understanding of the value such a team can provide.
We localized the app and website (and learned that sometimes this wasn’t enough)
Pipedrive translated the app and website to Spanish, German, Russian, Brazilian Portuguese, French and Estonian pretty early on. The effect of this varied greatly by country and language. In Brazil, for example, a prominent blog mentioned Pipedrive in a post around the same time, and this in combination with a localized app and website sparked growth. Where we failed to get local coverage, the impact of localization was more modest. See below: these are signups from a country where English is not widely spoken, where Pipedrive penetration was low, and where we didn’t do any local promotion after adding the translation. There was a bit of a signup increase, but this slowed down soon.
Localization without promoting it locally is like laying the fuse without lighting it. It seems Trello’s experience is remarkably similar.
We made Pipedrive talk to other apps
This one is difficult to measure, but anecdotal evidence suggests it would have been much more difficult to grow to 50,000 customers without having a fully functional open API early on. Having an open API allowed seriously tech-savvy customers built their own customizations and connections at a time where most other CRM apps didn’t have one. This created a small but very vocal group of fans at a time we needed them the most. And some of the tech-savvy customers that had begun to know and love Pipedrive have recommended us to 10+ other organizations.
Arguably, having an open API and a Zapier connection is table stakes rather than a competitive advantage today (similarly to having a mobile offering), but it’s still very wise to have one for several reasons.
We went beyond CAC in paid channel management
Pipedrive started experimenting with a 500€ per month Adwords budget as early as there was some spare funds available. (Literally not figuratively 500€, the early days weren’t all that glamorous). Despite having to turn off all paid advertising during one or two cash-strapped patches, by the time we had secured seed funding we had enough data to conclude that some keywords groups returned money in less than 6 months. We also learned that others that seemed equally attractive on cost-per-signup basis performed 2-3 times worse due to lower conversion to paid and/or higher churn.
Knowing what the LTV is from different channels and different ad groups has taught us which keywords groups, languages and countries to scale. This was confirmed a couple of months ago when we ran a similar analysis on a much larger data set.
We were creative with efficiency (or efficient with creativity)
It’s easy to be “creative” in a one or two-person marketing team. Want onboarding emails to generate some buzz in addition to showing people the best way to use core features? Easy – just take half a day more for writing. It makes sense to invest a lot of time in things like this because you desperately need their impact.
Controversially, the more resources we had, the less time we had for worshipping the devil in the detail. We had done this thing called “scaling” and were running increasingly large projects with increasingly efficient processes. So much so that we at times had scoped creativity out of our workflows. Some of our marketing had become efficiently dull. Several team members had come to similar conclusions around the same time and there have been efforts to get us back on the creative track ever since.
Creativity is something that’s been our strength most of the time in the early days and a lot of the times throughout the journey. It has and it hasn’t helped us to 50,000 customers. If I was more of a betting man I’d bet that it will greatly help to get us to next milestones.
— Maksymilian Robiński (@maksrobinski) February 3, 2017
We learned that ideas are worthless and that it’s all about the people.
Last but not least, the most important thing that got Pipedrive to 10,000 customers is our team. In our first couple of years, we in marketing were looking for ideas: a killer marketing campaign, a new platform or a press angle. And in the early days, this was enough. Over time two things happened. We got bigger and scalability of ideas got increasingly important. And we learned first hand that if you put the wrong person or team in charge of a great idea, nothing much happens results-wise. Or – you can have a mediocre idea or opportunity and if the right person or people are involved, it will turn into a goose that lays golden eggs (which is useful if you like a golden omelet).
The most important thing in getting us past the 50,000 paying customers mark, the real topic of this post, is our team. Our most important asset is not our tech stack or UX or, god forbid, marketing but our approach to hiring, onboarding, keeping and occasionally also parting ways with people.
Disclaimers and footnotes
And finally, one thing that has changed between getting to 10,000 and 50,000paying customers is my belief in posts like this. On the one hand, context and timing matters, and what works for X may work for X only. And on the other, I’m increasingly of the mind that there are several different good ways to market a software product such as Pipedrive. Many paths would have led to the aforementioned milestone, some potentially faster or easier.
If you’ve read all the way here, great. You’ve hopefully learned something useful and I’ve experienced a drop in bounce rate. A true win-win! I love to chat growth and marketing with smart people, time permitting, so feel free to get in touch via a comment below or shoot me an email.
Image credit: ostagram
Update November 2020: For newer readers, I quit Pipedrive in 2017 to work on Outfunnel, a revenue marketing automation tool.
Do you think that platforms like GetApp are still a good investment of time for a new Startup?
I’d say GetApp and other sites like GetApp are worth testing for most B2B apps
Would you say, with better fundamentals — you should’ve been doing this stuff from day 1 and would’ve hit 50k faster?
Not sure I understand what you mean by “better fundamentals”, Andrew.
Excellent write up. Thanks for sharing your learnings (congrats on hitting 50K!).
Regarding support with a human touch: what are your thoughts on using chat bots? AI continues to be an emerging trend in the SaaS world. Wondering what you think of it from a support standpoint. Thanks!
I personally haven’t worked with chat bots, and I’m yet to experience a great chatbot experience as a user / exploring marketer. But there’s surely potential there and I’ll keep looking. Any great ones you have seen?
Great article, Andrus.
Do you remember any rough timeline on how much time it took for you to turn the referral program into a successful channel? I spoke to a lot of SaaS marketers and most of them said referrals wasn’t working for them for some reason. Any insights from your experience of making a referral program work?
It’s seldom a rocket ship like it is so for Dropbox. But it can generate 2-10% additional growth in a very low-maintenance / high profitability manner. And you’d need a base of at least 1000 customers to make it worthwhile – the more the merrier thereafter.
Great article! You prove the point that neither marketing, l10n, customer support nor product management cannot work in silos – the overall goal and KPIs are the same.
Great insights, especially the one about the team. I recently started working with a great team and I must say – my ideas are getting better and more importantly – I’m executing them better.
The team drives me forward. Which makes sense, we’re all about doing great work and not just being stuck in a corporate structure. Keep up the good work! Looking forward to the 100,000 customers update.