To Be [CEO] Or Not To Be?
I didn't start out seeking to be CEO of anything; I came on as a software developer. The live chat company had no real website, for starters, and the employees were using a hodgepodge of disparate tools from across the internet to perform their day-to-day tasks. They needed an internal tool to bring everything together. The clients, also, had no ability to see or manage the language and content used by the chat representatives to represent them. They needed metrics and input.
I got to work, quickly building out the basics. Shortly thereafter, I was asked to serve as CEO, something I had never done before. I have a degree in Organizational Management, and a background in HR, so it was not completely outside my wheelhouse, but my current profession–and passion–is software development. I determined that it would be a challenge and an opportunity to learn, so I accepted, and I'm glad I did.
First, let's talk tech.
I got to work, quickly building out the basics. Shortly thereafter, I was asked to serve as CEO, something I had never done before. I have a degree in Organizational Management, and a background in HR, so it was not completely outside my wheelhouse, but my current profession–and passion–is software development. I determined that it would be a challenge and an opportunity to learn, so I accepted, and I'm glad I did.
First, let's talk tech.
The App
For the first few months, there was another developer, but he eventually went full-time as a consultant to the owners, and I became the primary dev. We built the app in Rails using the Jumpstart Pro template. I love this template for a bunch of reasons. For one, Chris Oliver is a great guy and he puts out quality, reliable content. For another, it includes a lot of sensible features you might want in a SaaS, like personal and team accounts, account impersonation for admins/troubleshooting, and payments using Stripe/Braintree/etc, among other things, and sticks closely to standard Rails practices.
Over time, the app included several things that I am proud of:
- Slack integration: Like many companies, the majority of our communication has taken place in Slack. This integration allowed us to receive notifications in specific channels when, for instance, a new client signed up or added their website, or a new demo call was scheduled. This helped us stay on top of things in a timely manner.
- Re:tune AI: re:tune may not be well-known, yet, but we have appreciated them. This integration has allowed us to train specific, separate, AI models on each client's website, in order to give our chat agents the opportunity to query against existing website content, allowing them to find and/or verify information before providing our clients' customer's with answers. One cool feature is that the AI stays updated when the website content changes. We were also able to add documents for each website/client containing information not publicly available.
- LiveChat Data: Under the hood, our agents have used LiveChat for handling incoming communications, be they from chat widgets on our clients' websites, SMS, or social media accounts. Our experience with this tool, however, has been love/hate. For the first 13 months, we experienced some frustrating things. A major annoyance was that while using the chat widget, the text input would frequently become disabled and a message displayed at the top advising the user that they had a poor internet connection. You could open other tabs and load websites, watch videos, etc, and therefore know it wasn't your internet connection, but this persisted. It often took a long time to finish typing a message. There were other things too numerous to mention here, but other major annoyances included the chat agents getting logged out of the software without knowing it, and not being able to log back in. Their uptime was pretty good… except when it wasn't. The last downtime almost caused us to switch to another platform: TWO AND A HALF HOURS they were down, at one time, without warning and without any explanation or apology. Why didn't we leave? Because somehow the other issues have decreased since then. This suggests that they were rolling out an upgrade that went south, which, though much appreciated, should have qualified the end users for the aforementioned explanation or apology. Sigh... Well, what I really wanted to say here was that we were able to integrate with their API and pull in data to display to our clients concerning the number of chats we had handled for them for both the current month and the last 30 days, as well as the number of leads submitted for the same time periods.
-
Easy, Consistent Onboarding: When I came on, the agents were using Notion for documentation. While Notion has its place, this was not it. It was way too easy for any agent to alter content, intentionally or not, and for the content to be very different from one client to another. When you are doing something like handling chats, consistency leads to speed because no matter which set of documentation the agents visit, they will know what they will see and where it will be. Less hunting = more finding. So, that's exactly what I built. After determining what kinds of information had been collected for our existing clients, I added a form for new clients to add their own site documentation. This included a process for the documentation to be approved both by the client and an internal admin.
- Disaster Notifications: Since our chat agents have all been located in the Philippines, we faced some challenges you are not likely to encounter elsewhere. For instance, one agent had to go offline for a brief period because her house flooded. Another lives in an area that frequently experienced earthquakes, often up to 6.0 on the Richter scale. Others live near an active volcano. One challenge posed by this situation was keeping a minimum number of agents online at a given time to ensure proper staffing and maintain our standard of answering all chats in under 60 seconds. To this end, I created a disaster notification system which included a Telegram integration that all agents could receive on their phones as soon as we were notified of an incident. The notification included a link back to the company app, where the agents were asked to self-report their current availability and the level to which they had been personally affected by the disaster. This data was used by the team lead, who then knew a) if anyone needed a substitute, and b) who was available. This, of course, also allowed us some insights into the personal lives of our agents, so we could all provide emotional and professional support during difficult times.
Obviously, the tech side wasn't the challenge for me. Now, let's talk business.
The Role
A CEO, at least in a small company, must be knowledgable about many things, such as business, finance, HR, sales, and marketing. I had experience in some of these things, but not others. We'll get into that later.
In order to understand where the business was positioned, I gathered data on the competition. There are a lot of options in the chat industry, and they all take a different approach to a number of things such as pricing and services offered. During these 14 months, the landscape also changed as two things happened: 1) AI chat got really good, and 2) people really started to hate it. #1 created a growing challenge for our small business, since companies saw it as a fast, easy way to implement chat. #2 worked in our favor because customers (ie, end users) loved what we were providing. The biggest challenge, and what ultimately led to our downfall, was that OUR customers–the ones making purchasing decisions–were those described in #1.
Ironically, our service was faster to implement than AI, super easy to get started with, and people liked it.
The other challenge that contributed to our downfall was the sales and marketing aspect. These were the two really important aspects of the position that I had no professional experience with and, unfortunately, little aptitude for. The owners brought on two salesmen, one after the other, who were both relatively inexperienced. They also both got hired out from under us by financial institutions, which is an odd coincidence. These gentlemen came in and attempted to use the sales outreach practices that have been successful for the past several years. What they didn't know was that Google and Yahoo were changing the rules at the exact same time so that those practices no longer worked. I just learned this myself last week. Had the top brass been willing to spend money on more experienced sales reps, we could have known about this as it took place (or before) and responded appropriately.
One final challenge that must be noted was that of having too many cooks in the kitchen. In the beginning, the owner, the consultant mentioned earlier, and myself, were all involved in day-to-day processes, dealing with agents, etc. This made it difficult to make decisions quickly, and often resulted in mixed signals, as one would occasionally make a decision that had not been discussed with the others. The turning point came when the other two left the chief executing to the Chief Executive Officer.
The Wins
When I took over, our team of chat agents was led by a woman who ruled by degrading agents, and lacked either compassion or understanding. There was no ongoing training. She had been with the company for, I think, 11 years. I eventually made the hard decision to replace her with an agents who had showed greater competency, and later let the former team lead go. This turned out to be one of the best decisions I could ever have made. The new woman was extremely reliable, and was effective at leading the other agents.
On a similar note, I should mention the time I let go of an agent whose wife was soon to give birth. This is not about a win, but a difficult correct decision. I had implemented a program for the agents to use their free time to earn some extra cash by assisting the sales team with their prospecting efforts. They weren't doing it and we needed sales badly, so I mandated that they submit a few names each week. The agents all personally committed to do so. Some of them still did not. In committing the agents to the task, I told them that they would be replaced if they failed to do it. I identified the poorest performing agent and let him go. Again, it wasn't easy, but it had to be done.
Now back to the wins:
A few months after taking the reins, one of our clients scheduled a video conference with me. They had had some bad experiences with our service up to that point, and wanted to tell me off before dumping us. I know this because they told me so during the call. Thankfully, I had begun implementing an improved training program, so that at the end of our chat the client asked me what we were doing differently. Right before our call, he had gone over the more recent chats and seen the improvement. They decided to stay, after all.
I continued focusing on improving our quality of service by improving our training program, and it has paid off. We have the same agents, but now, 14 months later, they get tons of positive ratings every single day. Our clients' customers love them, and I couldn't be more proud.
The Takeaways
1. Let your leaders lead. If you trust them, get out of their way. If you don't, get rid of them!
2. Don't scrimp on your sales team. You get what you pay for.
3. It's always worth it to invest in your team. They fight your battles, give them the best weapons.
4. An ordered world is a better world. Take the time to organize and simplify your documentation and processes. Create SOPs.
5. Do the hard things, even when they feel mean. Your team needs to know they can rely on you, including the things you say.
2. Don't scrimp on your sales team. You get what you pay for.
3. It's always worth it to invest in your team. They fight your battles, give them the best weapons.
4. An ordered world is a better world. Take the time to organize and simplify your documentation and processes. Create SOPs.
5. Do the hard things, even when they feel mean. Your team needs to know they can rely on you, including the things you say.