How we failed our first OKR cycle and why we do virtual teams now
When a company is just 10 people, you can get away by delegating tasks directly. But it fails with a loud bang when you grow to 30+ people and some company structure starts to appear. How do you align the teams? How do you structure common goals? How do you track those goals?
OKR framework is a very likely answer to all of those questions. It is a popular and simple approach that is even successful at some companies. We switched to OKRs last year and since then, I have witnessed 4 OKR cycles and led 2 of them myself. Of course, we made mistakes, learned, and iterated, as all startups do. Here is how it all went.
OKR cycle 1
So, we are doing OKRs! Such a glorious moment. We found a fitting software called Perdoo and at least 50% of our employees were excited about the future. Good enough for a start.
As in many companies, we appointed our HR manager to be an OKR ambassador. Then the leadership team took our quarterly goals and translated them into objectives and shared them with the managers.
We had six core teams that roughly got the following objectives (I edited them, but they generally illustrate the point):
- Product: release deal creation manager and a new back-office tool
- Tech: setup data warehouse and hook it with Tableau
- Marketing: improve organic traffic conversion by +5%
- Operations: reduce credit check time by 50%
- Finance: automate 50% of invoicing and payment actions
- Sales: get 400 new customers and decrease time to deal down to 10 days
As you noticed, half of the objectives are features and not goals. But that was not the biggest problem, everyone goes through that. What we haven’t seen is the bottleneck we have created.
Lesson 1: Don’t define OKRs without teams input
It now sounds very stupid, but after 1 month into the cycle, we realized that most of the teams needed tech and product help to achieve their objectives. Finance could do some basic automation in Google Sheets with Zapier, but the real boost required a tech solution. As was the case with most of the teams, unfortunately. Not only that was a disaster with many overtime hours ahead, but our investment tranche was tied down to those OKRs. Oh, how naive we were!
Thanks to the hard work of the team, we achieved most of the OKRs by brutally reducing the scope, cutting all the possible corners, and releasing nano-MVPs with fake doors and mechanical turks. But I swore that I would never do that again because I can’t eat so much pizza for dinner in the office.
Protip: Best OKRs are small and respect your limited resources.
Lesson 2: Don’t let people with no time drive OKRs
In our case, the founders were busy managing investors, planning the strategy, and securing partnerships. HR, no matter how well they were connected to every team, had very little free time & business overview to plan an entire quarter for 30 people. They can support you in the process, but someone else should be the driving force. Ideally, one of the founders, or at least a C-level.
Since VAI is a product-driven company, it made sense for the product team to get into the driving seat. Luckily, I was both a CPO and passionate about OKRs, so I volunteered to organize the next cycle and make sure teams didn’t overload each other with dependencies.
Protip: Ideal OKR driver should possess time, seniority, passion, and vision.
OKR cycle 2
Our HR went into maternity leave, and since our previous attempt has miserably failed, we wanted (especially me) to do it properly this time. We agreed to skip a quarter, clean up tails in Perdoo, and focus on the next one instead.
OKR cycle 3
So I set down with each team and explained my plan. Let’s take company goals, look at the numbers we want to achieve, and brainstorm how each team can contribute. But. Each team should either solely rely on their resources or book tech and product teams in advance. No more doing the same mistake twice.
And that worked pretty well for us. We have hit most of the key results and OKRs were back in play.
Lesson 3: Don’t let two teams do the same objective
For example, both the sales and marketing teams set a goal to reactivate dormant customers by reducing their number from 600 down to 100. Sales called the customers and offered discounts, marketing has sent emails and educated about the recent changes in the product. The list quickly dwindled but no one knew which team made the most significant dent.
On the one hand, we could have split objectives and measured them in isolation. Sales could track the number or reached customers that in the end, submitted a deal; marketing could focus on the open rate and newsletter conversion. On the other hand, that would make the objectives smaller and less attractive. And goals are supposed to look big, ambitious, and tasty.
Protip: Collaboration works better than the competition in OKRs.
Lesson 4: Not every team should have an OKR
The operations team was reducing the time it takes to: review an invoice, approve a deal, pay the supplier. However, they soon faced a burden of reporting their progress.
The reason was quite simple — this objective was not solving a real problem. We don’t have 1000 deals per day to approve, reviewing invoices was never a bottleneck; neither were payouts.
When you are a small team that mainly does daily business, like operations or recruitment, most of the OKRs would feel either far-fetched or forced. So we agreed that some teams would not have objectives until it’s time. When our Ops team would grow to 10+ people, then we will start thinking about saving minutes on repetitive tasks.
Protip: Doing OKRs for the sake of OKRs greatly lowers morale.
OKR cycle 4
This cycle has not started yet, but we still have unanswered questions from the last quarter. Some of the teams are working on the same OKRs; some don’t have OKRs at all.
To address that, we will introduce virtual teams. A team that works on a chosen OKR but has members from different departments.
As usual, we will post top-level company goals in Perdoo and give everyone a week to come up with supporting objectives. However this time, each objective will have a champion who will pitch it. We expect no more than 15 ideas (we are 40 people) and the champions will most likely be our team leads and managers.
At the end of the pitches, every person in the company can choose up to three objectives that strike him or her the most and join their virtual OKR teams. The leadership team will then prioritize the most relevant ideas and kick off the new quarter.
The idea champion is the main driving force who executes on objectives, organizes team meetings, manages stakeholders, and reports the progress. This works only in a relatively flat structure where people don’t ask for permissions but rather report results.
In this scenario, we expect more cross-team collaboration and better idea ownership. Of course, some people will not join any teams and will focus on daily business such as supporting deal flow in operations or doing user research in the product team. And that’s perfectly fine too.
Protip: OKRs should take around 50% of the time of approximately 50% of the people.
Lesson 5: Not learned yet, but …
In most companies, and VAI is not an exception, OKR is a very top-down approach. Yes, teams are involved, but they have boundaries set by either the business plan, investor expectations, or top management vision.
If virtual teams solve our cross-department collaboration issues, we will invite our idea champions to suggest quarterly goals for the company. So that we have a mix of both strategic objectives derived from the vision and down-to-earth objectives raised by the teams. We have so many bright people here I am sure they have a lot of insights that leadership is not seeing.
Daniil Pavliuchkov is a product consultant and speaker helping companies grow by improving their OKR and product processes.
Upcoming talks: “How to grow your PM team into heroes” at the Product Management Festival in Zurich, 13th of November. (Ping me for a 20% discount)
Illustrations by Viktoria Kurpas.
