Build A Strong Middle Management

Managers backlit

One of the most important elements of scaling a business is you must develop a strong middle management layer. Often when a company first starts to expand, people are made managers somewhat haphazardly. Perhaps they are the least worst for the job, the most outspoken, the first hired, the tallest or some other random criteria. Sometimes people become managers without ever being told what that entails. Some figure it out. Many do not.

Yet, the middle management layer is the key to scaling a company. If you don’t have a solid front line management and second level directors, you are unlikely to build a high-performance culture, no matter how good your executive team is.

Why is that? At a fundamental level, the middle managers outnumber the executives. They also have more direct impact on the work of their teams. So the span of influence of the middle managers is greater than that of the executives. You can come up with all the great ideas and strategies at the executive level, but without buy-in and leadership from front line managers, very little will get done.

So it is critical that you develop the performance of the middle managers. Their influence and abilities will determine whether your company will scale to great heights or whether you will be a company with brilliant strategy and lousy execution.

The guiding principal should be to push as much decision making as you can further down into the organization, close to where the action is. The idea is that those closest to the problem usually have the best context on how to solve the problem. You want to empower front line managers to make the decisions on what work gets done, how to prioritize and when to make exceptions.

But, wait a second — how can you trust these inexperienced managers to make decisions?

It helps if you have a meaningful set of values and principals that guide your operations. These can’t platitudes. They must be guidelines that will actually help people.

As an example, at Gatsby one of our values was to prioritize the customer. But we didn’t just talk about it, we lived it. We would go the extra mile to help customers out when they were going live into production with our product. Sometimes that meant 5am calls with our engineers and managers working side by side with customers. It won the customer’s trust and support. It also demonstrated inside the company, that we were serious when it came to providing great service. And if a customer was unhappy with the service, for whatever reason, we would let them out of their contract. And it wasn’t me making these decisions. It was the front-line managers.

It isn't easy to be customer focused. But over time, we turned it into a competitive advantage. We often won customers from competitors because they felt screwed-over or trapped into contracts that left them angry. Everyone in the company understood the competitive value of helping our customers. They saw that the managers and executives focused on customers, so they did, too. We gave examples in our regular town hall meetings. Values like customer focus were part of our annual review process. When someone was promoted, we showed how they helped customers. This encouraged employees to take initiative themselves, without having to ask for permission.

At Zendesk, I remember one time Mikkel the CEO asked me “Why do I have a line outside my door every morning?” It was because he was letting people escalate decisions to him. With every decision Mikkel made, he was reinforcing the behavior, instead of encouraging people to make decision themselves.

I never like it when people bring me decisions that they should handle themselves. It makes me anxious because they likely have more context than I do. And often, by the time the decision gets to me it’s become quite a thorny customer situation or a struggle between two different departments. If I can, I will ask questions to elicit more information, discuss the tradeoffs and encourage the manager to make the decision based on our values.

Scott, our director of Finance at Zendesk came to me one time and asked me to implement a discount approval schedule, along the lines of sales reps could approve up to a 10% discount, managers 20%, directors 30% and perhaps the VP up to 40%. Anything beyond that would require my approval.

I said no.

Scott was stunned. “What do you mean?” he stammered.

I told him we didn’t need a discount schedule. I trusted the sales VPs and their managers to implement whatever discounts they felt were necessary.

“So you’re saying if Matt wants to discount the product to 90% you are ok with that?”  (Matt ran our European operations and was one of our most trusted executives.)

“Yes,” I said. “Matt is an experienced guy. He’s not going to discount something if he doesn’t need to. He’s also not likely to do anything radical without discussing it with me. He runs his region and he knows his customers and the market in Europe better than we do. So if he needs to provide a discount, he’s the best person to make that decision.”

Scott walked off infuriated. He was used to a command-and-control model where headquarters made the rules and slapped people down when they got out of line. I understood where Scott was coming from. After all, that was the model at every large software company, including Salesforce where Scott had worked previously.

But the role of an executive is not in approving things. The goal is to add value. The best way to do that is to create a resilient, high-performance culture. One of the ways to do that is to empower your front line managers and executives and teach them to make good decisions. If people want my input, I provide it. But I don’t hire people so that I can second-guess them. I want managers to grow in confidence so that they are making good decisions, developing action plans and communicating.

In my experience pushing decision making close to the action is the best way to operate. Most people will think things through when it’s their call. I want to cultivate an organization so that when the stuff hits the fan, managers don’t panic and instead they muster their problem solving and communications skills to tackle the problem. And they can’t do that if they’ve never been given the responsibility and trust to make decisions.

How does your organization make decisions? What decisions come your way that could be handled by someone closer to the issue?

Setting and Measuring Objectives


One of the things that I see routinely in early stage startup companies is a lack of accountability around execution. Sometimes this is because there are unclear objectives. Or sometimes there’s a process around objectives which has become dysfunctional or fallen by the wayside.

When I joined Duo Security, the founders Dug and Jon were proud of having an OKR (Objectives / Key-Results) system in place. But as I went around the company to speak to various managers, it became evident there were problems. People wrote down all kinds of ambitious objectives at the beginning of the quarter, but then no one looked at them again until the quarter was over. There was a perfunctory assessment of the results, they were shared in an all-hands meeting and life went on, un-impeded by the objectives. 

OKRs were originally developed at Intel and then became more widespread in the startup world after Google used them. Unfortunately, somewhere along the way, OKRs were corrupted. One of the main problems that I discovered with startups using OKRs was the belief that objectives should be extremely ambitious (also known as “stretch goals.”) Therefore, if you only got 70% of your objectives that was considered a good outcome. I am not sure if there are other disciplines where 70% attainment is graded as 100%. 

Unfortunately, this is a primary source of OKR dysfunctionality. Because it can lead to an almost random de-prioritization or dropping of 30% of the objectives. What happens if the Engineering team commits to an ambitious set of features to be delivered in the quarter and then mid-way through the quarter drops 30% of them? What if Marketing spent the first month of the quarter working on launching those capabilities and instead dropped some other features from their list? And what if Sales has been training people on new features that no longer exist? 

It becomes quite clear that by being overly ambitious, we have created a lot of thrashing across departments. The intersection of what needs to be done is hampered by sub-optimal decisions made at a departmental level. 

The solution at Duo was to create a less ambiguous process and brings a higher-level of clarity and coordination across departments. It begins by having a small but clear set of objectives at the corporate level. In an early stage company this likely includes things like:

  • Launch new Enterprise edition by July 15

  • Expand into EMEA with translated product in German, French

  • Improve reliability by having no unplanned downtime 

  • Raise Net Promoter Score (NPS) to 30 or higher

  • Generate 1,000 MQLs and sales pipeline of $2m

  • Generate $1m in new bookings

Each department that worked for me (which was initially sales, marketing, product, customer service) would create their departmental objectives, taking into account the overall company objectives. However, we were not obsessed about forcing departmental objectives to line up with specific corporate objectives. Sometimes a departmental objective is just its own thing. Worrying too much about which initiative departmental goals align to can obscure the more fundamental decision: is it worth doing or not? 

Although there are lots of different tools that you can use for writing and managing goals, in my experience the tools don’t matter. (And in fact, fancy tools with too many capabilities can add more confusion, for example surfacing debate about how to weight various objectives.)

I like to keep things simple.  Write out the top 8 or 10 or 12 objectives in a text document in an enumerated list. I weight them all equally. So if one goal is substantially bigger or harder than others, consider breaking it into two. The objectives should be SMART, e.g. Specific, Measurable, Attainable, Realistic, Time-based. Most fundamentally, at the end of the quarter, we should be able to determine whether the goal was completed or not. I like to make sure that executives have skin in the game so we tied half of executives bonus to be paid based on company objectives (usually oriented around the Annual Recurring Revenue, or Sales bookings) and half based on attainment of their departmental objectives.  

I also make it clear that the executives must coordinate with their colleagues across departments. So, for example, the head of sales and the head of marketing must be in agreement about what are the appropriate objectives around lead generation, sales tools, training and so on. It helps to reinforce the “client” relationship that exists between departments.  That is, I want the head of marketing and the head of sales to recognize that a big part of the marketing department’s role is to help sales close more, bigger deals faster. I want them to work out these things together rather than throwing it over the wall, or worse, sending it up the organization to the c-level executives to determine the appropriate trade-offs. If the heads of two departments can’t work together on the same side of the table to prioritize goals, then its unlikely that people in their departments will work together either. The executives must set the standard for collaboration. (If they can’t, they probably shouldn’t be in the company.) 

My direct reports are usually fairly senior, so I tell them if they get 80% or higher of their objectives, life is good and their bonus will reflect their accomplishment. At 70% it means we did not set the right objectives or I did not manage them well enough, so I will manage them more closely. At 60% or below, they can’t work for me. Since I also say this in the interview process, it helps candidates who are not goal-oriented or do not want to be held accountable to opt out of the process. 

However, I try to be flexible in one respect. I tell people that they can come to me any time during the quarter and if there’s something they set as a goal that they view as no longer useful to the company, we can strike it from the list. However, we don’t drop something because it is hard, only because it’s no longer the right priority. In my experience, it’s fairly rare for people to take me up on this, but it prevents people from working on something they realize is no longer important. I also remind them that they cannot come to me at the end of the quarter and say the accomplished a whole bunch of other work instead of what we agreed to.

It’s good to follow up with people mid-quarter or in some cases, monthly, to review their progress on the goals. This is an important reminder process sometimes things have either not started or we’re not yet seeing good results. It also helps frame a conversation around the objectives, what is blocking them, and how we can increase the likelihood of success. 

Setting objectives at the company level and even the departmental level is not easy. It requires focus and discipline. Often high-growth companies are filled with people who are ambitious and want to do lots of initiatives. Most organizations in a growth phase benefit from doing fewer things better rather than being spread too thin on too many things. You have to say no to some things. 

Of course, there are many more things that get done in a department than the eight or ten top objectives. But it helps to identify what are the most important items that will move the company forward. 

I have found that if you roll this kind of framework for setting, measuring and rewarding objectives at the top level after one or two quarters, it can be rolled out more broadly, to the next level of management. I would not suggest rolling it out out across the entire company all at once. When rolled out too quickly, it can easily lead to chaos at the end of the quarter with employees chasing down their colleagues in other departments to do work in pursuit of “their” objectives. It is better to take a measured approach. Get good at setting objectives at the departmental level before you expand broadly.

How does your company set goals? Does everyone know the overall company goals for the next quarter? How often do you follow up on the goals? 

The Cost of Poor Performers

Thumbs down 2
It’s always tough to deal with poor performers. No one likes to manage laggards or to admit to a hiring mistake. When companies are growing at a fast rate (50% year over year or higher), managers will often focus more on recruiting new people than in managing the team they have.

So what happens?

Inevitably, as a company hires beyond its network (of known good people) there will be some number of bad hires. Maybe they bluffed their way through the interview process, or no one checked references.  Sixty or ninety days out, when someone asks the manager how new guy Fred is performing the answer is “the jury’s out.” 

This is nonsense.

If you can’t tell what impact a new employee has had in their first couple of months, it means there isn’t any and in all likelihood there won’t be much improvement down the line. Because if you hired someone who was good at their job (or ideally excellent) you would already be seeing the results immediately.

So now you have the pain of managing someone out. Put together a performance improvement plan, give them a second chance, if it makes sense and get them up or out.

Actually, it’s worse than that.

In any organization that’s growing quickly, once you get beyond about 30-50 people it’s likely there are some people who are not quite getting the job done. Everyone deserves the benefit of the doubt and feedback on how they can improve. This is especially important in an early stage company where much of the culture is implicit and not necessarily documented or understood by newcomers.

In most cases, poor performers are doing their best to just survive and put up a good front. If managers are busy, they may assume their new hires are getting things done and not even notice any problems.

But the new employee's colleagues, with whom they spend more time, are much more aware of who is getting the job done and who isn’t. They might be inclined to help a new employee come up to speed, but they have their own work to worry about. If left too long, the top performers will wonder why no one is taking action with the poor performers.

Few people like to grouse about a peer’s shortcomings. But they all see it. If you’re in an office, they will see it and hear it. In a distributed environment, it will be harder to notice, but sharp observers and hard workers will wonder why Fred is taking so long to do something that they could have handled in a few hours.

If you’ve worked with great people in a high-performance culture, your tolerance for mediocre performers, bureaucrats and talkers starts to fall. And once the number of poor performers gets to a certain level, the best people will leave. If management can’t  tell the difference between good and bad performers, why would they stay? And when you lose top performers, you lose the creativity and drive that is essential in building a great company.

You need to give clear feedback to employees when they are meeting objectives and when they are not. You can’t wait until the end of the quarter or worse, until the annual review process. Set clear expectations: here’s what you need to deliver in order to continue working here. Follow up in writing. Help them improve. But remember, as a manager, you are the coach. You have a stopwatch, you set the goal, but they must run the mile.

I have a very high expectation of employees. I want everyone to be excellent in their job, whether they are a product manager, a programmer, a sales person or an administrative assistant. So the two questions I ask are:

  • Are they excellent?
  • If not, can they become excellent?

If they can’t become excellent, I would rather cut them loose and invest in someone else with higher potential. In all the years I have helped managers with these decisions, no one has ever come back and said “I wish I’d given them more time.” Usually, it’s the opposite. They wish they’d moved faster so it was less of a drag on them and the team. The good news is, when you do make a change, everyone’s productivity goes up. And when you let someone go, treat them with respect and help them leave with their head held high. 

How do you evaluate the performance of people in your company? How often do you give feedback about what is expected and when? What happens when someone doesn’t get the job done? Do you give them a second chance? A third? A fourth? Do you wait until the end of the quarter to give feedback? Why?

What Are The Patterns?


One of the questions I often ask founders is “What is working and what’s not working in the business today?” What I’m trying to do is identify the patterns they see among customers.

Identifying patterns and reinforcing them is one of the most important elements in building and scaling a startup. Unfortunately, it’s not a common skill, and if you don’t have it yourself, you will want to make sure it exists somewhere on your executive team.

In an early stage startup, it is essential to to seek out patterns among users and prospects to identify those with a higher likelihood of paying you. The patterns are key elements of why your product resonates with certain people and not others.

While it’s tempting to make this a quantitative exercise based on title, company size, vertical market and so on, the question you are trying to answer is: what causes some prospects to buy and not others? Demographic issues have an influence, but it may also come down to the particular problem the customers are trying to solve and the circumstances around it. I have found it helpful to understand the qualitative elements around customer situations and then try to determine ways to enable customers in those situations to put up their hand and identify their need.

At MySQL, we had a scale problem many startup companies would envy. We had millions of users and thirty thousand new downloads every day. The high volume of users made it harder to distinguish between free open source users and corporate customers who might be inclined to pay. A big part of the job in marketing and sales was to identify patterns which put people into one bucket or the other. We developed a content strategy that made it easy for users to self-select into different categories.

For example, we knew that there was a class of applications for embedding a SQL database into an application or hardware appliance for which MySQL was demonstrably better than competing databases from Oracle, Microsoft and others. So we created webinars, white papers and case studies that highlighted how to evaluate MySQL as an embedded solution. These were targeted specifically to engineering managers and product managers who were the most likely decision makers. Now instead of searching for a needle in a haystack, we had a much more clearly set of prospects who were trying to learn all the could about the problem they were trying to solve.

The embedded use of MySQL wasn’t the dominant pattern among our users. It was probably somewhere between 15 and 20%. But we knew they had very specific requirements of low-maintenance, high efficiency, low-memory requirements for which MySQL was ideally suited. And it was a very profitable segment with deal sizes that could be well into six figures.

We developed similar content strategies for CIOs (and those who wanted to become CIOs!) to learn about open source technology, for Database Administrators who wanted to improve SQL performance, for users concerned with security and so on.

Once you uncover patterns about your users, you can use that information to better optimize your sales and marketing approach. These patterns may also suggest opportunities for features which better serve specific segments of the market.

You don’t need a pattern to represent the vast majority of your buyers for it to be useful. As long as it can be identified and targeted, it may be worthwhile. (Or said differently, there is rarely a single unified pattern that meaningfully defines all your users.)  In fact, trying paint such a broad picture of your audience may yield such generic information as to be completely useless. It is better to be laser focused on specific types of customers and use cases than to be a wandering generality.

And although quantitative analysis can be helpful, I have found that the two best ways to uncovering patterns are to talk to customers and talk to sales people. Just ask: why did they buy from you? And listen.  The most successful sales people often have a highly developed intuition around customer patterns. But in my experience, only about 20% of sales people are good at explaining those patterns back to management. So your best bet is to do customer visits with sales people and then ask the sales rep how often they see similar situations. That’s also a good reason to have marketing people take part in customer calls. You want your marketing team seeing what the sales people see and helping to identify patterns. 

What experiments can you run to uncover patterns among your users? What makes some users buy and others stall? And how can you better serve those distinct segments? What changes to your product, marketing or sales process can you use to optimize for patterns?

Are there patterns you've seen in your business that have been helpful? Let me know by posting a comment below. 

The Power of Velocity


Running a startup can sometimes feel like hell. And as Churchill said, if you’re going through hell, keep going. Forward progress can be hard to maintain when there are customer disasters, technical setbacks, funding challenges and more. But what other option is there?

As a startup, making forward progress is essential. And if there is any competitive advantage a startup can have over larger competitors, it’s that it can and must move faster.

Let’s face it, most large companies, no matter how agile they once were, get bogged down. The more success they have, the more they grow, the slower they get. Middle managers are hired, rules are made, and new processes are created all in an effort to ensure that no mistakes are made. But those processes and the bureaucrats who create them also cause a side-effect of slowing things down. The culture that once boasted of moving quickly, becomes complacent and accepts that because there are more people, things take longer.

New competition? Let’s take a wait-and-see approach. Sales problems? Schedule a review at the end of the quarter. Need more engineers? Budgets have to be approved. New product idea? Surely that can wait.

Most startups move fast. But if you’re not careful, you can end up re-creating the same bureaucracy found in larger companies.

If you’re the founder, CEO or executive, you can’t make all the decisions youself. That will only result in slowing things down as a pile-up forms outside your office, your email in-box or your Slack DMs. So you must create a culture that empowers others to make decisions. And you must set the pace for making decisions quickly.

In an efficient startup that means if a problem is found on a Monday, you’re brainstorming ideas on Tuesday, picking the best solution by Wednesday, implementing it Thursday and measuring the results by Friday. That might seem optimistic for some problems. Engineering complex solutions might take weeks. But for many other problems in sales, marketing, customer service or finance, it should be achievable.

Startup leaders need to demonstrate the importance of velocity to their teams. There will always be a desire for more more input, more discussion, more research, more data, more analysis. But that must always be tempered against eroding the advantage of startup speed. And in truth, how often does delaying a decision yield better results?

In the early days at Zendesk, we faced significant competition from Salesforce who had acquired a smaller competitor of ours called Assistly. Assistly was priced cheaper than we were and appealed to very cost-conscious small businesses. We created a “never lose on price” campaign which gave our sales managers the ability to aggressively compete on price against Assistly and other low-end competitors. We trusted our sales team to not lower their price unless they had to. And we knew that by approving a price match in real time with the customer demonstrated our ability to respond quickly to their needs. We also knew that getting a similar discount app, a trial period extension, or other change from Salesforce would typically require three levels of escalation taking weeks inside of Salesforce.

It’s possible that we left some money on the table with customers who might have gone with Zendesk instead of the lower-priced rival anyways. But the confidence the sales team developed during this process, the momentum within the company and impact on the culture was well worth it.

As you grow your company, you will often need to develop additional layers of process. There will be board meetings, executive meetings, departmental review meetings, pricing committees, product councils, launch meetings, and a myriad of short-term task forces. In all cases, these processes will be set up to generate better decisions, or more transparency, or higher collaboration, fewer surprises.

However, if you’re not careful, meetings and processes may end up bogging down decision making and watering down accountability. So whenever you define a new process, be clear on who owns the decision and when it will be made. And at any point if you are reaching the point of diminishing returns from these meetings, ask that the decision gets made immediately.

If a process or meeting has outlived its usefulness, don’t be afraid to disband it. This is especially important as you bring in new managers or executives. Give them the opportunity to eliminate the bureaucratic processes of their predecessors by paring back the number of attendees or scope of meetings.

As you look across your organization, are there decisions now taking longer than they should? Are there too many layers of approval? Is there confusion about who makes the decision? What steps can you take to cut through the layers so that speed remains a competitive advantage?

Let me know your thoughts by posting a comment below.