As a part-time entrepreneur, how can one achieve an ARR of $310,000 for an AI+Excel project within half a year?

In today’s globalized world of SaaS and AI, grand narratives are no longer the only voices. A small team or even one person has the opportunity to create a legend. One-person companies may become the norm in the future. David Bressler is the founder of formulabot (originally named Excel formula bot), which can quickly convert text instructions into Excel formulas through AI and provide corresponding Excel and Google Sheets plug-ins.

Formulabot provides free services, but each user can only generate formulas 5 times per month. If you need unlimited generation and other advanced features, you need to pay $9-15/month.

Since its launch in September 2022, formulabot’s MRR has exceeded $26,000 in 6 months, and the official website currently has more than 300,000 visits per month. This number may not seem amazing, but if I tell you that this was achieved by the founder David Bressler alone, in his spare time, and with only $5,000 in total marketing investment, what do you think?

The birth of creativity

David’s main job is the head of the data analysis department of a company. In May 2022, his second child was born. During his paternity leave, he noticed that AIGC was developing rapidly and truly entered people’s lives.

The idea for the startup originated from a basic Excel question David asked GPT : “What is the Excel formula for calculating the sum of column A?” After just a few seconds, he got the correct answer. But as he asked more complex questions, David found that GPT’s accuracy was about 85% – although this accuracy was good enough in most cases, it was far from enough in the field of data/analysis.

David saw an opportunity. Given his strong background in analytics and Excel, he knew he could improve the model’s output and build a more accurate AI model. Later that day, he searched online to see if there were any AI-based Excel formula builders on the market, but he didn’t find any, which strengthened his confidence.

Random question: If you discovered this demand and there was no corresponding product on the market, what would you do?

a. You definitely won’t be able to make any money from such a small product, and it would be a waste of time to make it, so forget it.

b. That’s quite interesting. I need to conduct extensive research and interview a large number of users to understand their needs.

c. I am too familiar with the industry and have been working in it for more than ten years. I can start directly without any verification!

Based on past experience, David thought he didn’t need to further verify PMF: he had worked in the marketing analysis field for more than 12 years and used Excel every day; considering that there were 1 billion Excel users worldwide, David thought the market was large enough; even in the worst case, he could build a product for himself and greatly improve his work efficiency. So he decided to quickly build his first version of the product.

Building an MVP

David believes that in the field of AI, industry pioneers will have advantages due to data and other reasons. He also understands that he is not the only one who realizes the great potential of Excel+AI, so he decided to act quickly.

However, David faced a serious problem at the start – as a data expert, he had almost no experience in building websites, let alone developing a SaaS product with AI capabilities.

Random question: If you encounter this problem on the first day of starting a business, what would you do?

a. Quit my job, find a VC, build a team, and I will be the next Ivan Zhao!

b. Find some brothers who know how to build websites and do it part-time!

c. I want to be self-taught, so building a website shouldn’t be difficult, right?

David chose option d: he created an app with only the most basic functionality, with no payment, login, or other complex settings. The app only contains a field for input prompt, a button to generate the formula, a field to display the output formula, and a button to copy the formula (as shown in the figure below).

David shared this product on the Excel subreddit and received a lot of feedback and suggestions from users, which were later used to improve the product; it was also shared to several other Reddit communities and even led to viral fission on social media – it also received a lot of attention on TikTok, Instagram, and Twitter.

David was very excited about the popularity on social media, but the Open AI API bill that followed made him collapse – since there was no payment and login function, he had to bear huge AI costs. This made him realize that he had to build a real SaaS product and start charging.

Random question: How would David configure the first payment feature?

a. Register a company and connect to Stripe. Do it the formal way.

b. Find some low-threshold charging channels such as lemonsqueezy and start running first.

c. Is it okay to ask users for tips?

However, website construction cannot be completed overnight, so David came up with an idea – he quickly set up a “reward” page on the official website , allowing users to give him a little bit of sponsorship. Unexpectedly, he later received thousands of dollars in tips from hundreds of users!

Officially released, self-built official website

It took David several months to develop the MVP into a complete website that supports subscriptions. The new website included login and payment functions, allowing subscribers to generate formulas unlimitedly, but David began to worry: “Will people really pay for this?”

Excitingly, the new website received its first order within five minutes of launch!

In the first week of the product’s official launch, David also did a Launch on ProductHunt and offered a big discount, which also brought him several thousand dollars in revenue and gave him further confidence.

So, how was the website built?

David is self-taught! However, this is entirely based on the rich “software supply chain” overseas. He built the official website through the code-free tool Bubble.io, without writing a single line of code for the entire website. The previous single-page MVP was also built by him in a few weeks through Bubble. The “software supply chain” overseas is extremely rich. Entrepreneurs only need to develop the core part of their own products, and the rest can be solved through suppliers. If you encounter difficulties, you will most likely find corresponding tutorials through forums or YouTube.

In addition, for David, management costs were very important. He did not raise funds or borrow debt, and he operated the entire business with his own funds until the website started to make a profit. In order to ensure cash flow, he did not even quit his job and developed his startup only in his spare time.

Ways to attract and retain customers

Seven months after launching, David’s MRR has exceeded $14,000 and continues to grow, recently reaching $22,000. Yet, his total marketing spend is only $5,000! It ’s important for a product to inspire users to share, but another important factor in this case is KOL marketing.

As a data expert, David was not optimistic about KOL marketing at first because its impact was difficult to measure and quantify. However, most of the traffic that formulabot gained in the early days benefited from KOLs in TikTok. David found that once a big V started talking about something, many mid-level KOLs would follow suit and discuss the same topic the next day.

However, since most KOLs mainly promote physical goods, David did not find a suitable partner after searching for many KOL Marketplaces. He decided to build a SaaS-focused Marketplace himself.

David’s strategy was to directly contact KOLs whose followers matched his user profile, who would either be willing to promote the product for free or for a fee (usually between $200 and $3,000).

David used TikTok’s Creator Marketplace to locate potential partners. It was completely free and allowed him to filter by keywords, number of followers, and other dimensions. David found that the number of followers was not positively correlated with conversion rate, and he would focus on finding KOLs whose followers were located in the United States, the United Kingdom, or Canada, as these regions are the main sources of paying customers. In addition, he also tried to find KOLs in specific industries, such as accounting, finance, data analysis, and other fields.

David has achieved a certain coverage on social media through KOLs on TikTok and Instagram; he also actively participates in discussions on Excel topics on Twitter. Here is a little trick: David sets up message alerts for the KOLs he focuses on on Twitter. Once he finds their tweets, he immediately comments on them. This allows him to occupy the first comment most of the time and also brings him a lot of traffic.

David basically relies on others to promote the website, which allows him to focus more on customer retention. David established a customer journey and pushed personalized information through multi-touchpoint automated emails: each message that reaches the user is differentiated based on many factors such as when they created an account, whether they paid, personal email or company email, etc.

Email remarketing brought in a fair amount of traffic, but more traffic came from word-of-mouth from colleagues, friends, and KOLs: organic search and direct traffic continued to make up the majority of website traffic, with the main Google keyword being “excel formula bot” (formula bot was originally named “Excel formula bot”) – which highlights the value of choosing an easy-to-remember brand name.

Random question: Which keywords do you think this product should use for SEO?

a.Excel

b.AI Excel tool

c. Specific function name of Excel

SEO is also important, and David has been building content pages around keywords that people search for specific formulas, such as “VLOOKUP” and “SUMIFS .” formulabot ranks well organically for these types of searches and has already received high-quality backlinks from sites like pcworld.com, inc.com, wellandgood.com, and BGR.com.

Although David had never done SEO in his full-time job, his team had always had an SEO leader, so he was familiar with SEO. At the beginning, David’s method was simple, which was to add popular keywords in the SEO tag and copy them throughout the website, and the referral links kept pouring in (there was no link building strategy).

An interesting thing is that since formulabot is the first AI Excel website, and the brand name is also the category (“Excel formula bot”), Google can see a lot of searches like “Excel AI generator” and “Excel formula creator”. These searches did not exist before formulabot appeared, which means they were looking for formulabot but didn’t know the exact name! Therefore, these keywords, along with many other searches, are included in the SEO title and meta description of formulabot, as well as other parts of the website!

Future plans

Although it started out in a loss-making state, formulabot is now profitable and growing. With the increase in subscriptions, cost optimization, and the development of other partners (such as the Excel course website and the Excel template website), formulabot’s gross profit is now increasing every month.

Try paid advertising

Since SEO is not a quick-acting method, David also started to invest in pay-per-click PPC ads, mainly targeting people who search for Excel formulas or Excel courses. The cost of getting each lead is still relatively high, and considering the LTV and paid conversion rate, the ROI has not yet turned positive. But David decided to continue investing because he believed that although these potential customers did not pay, they might tell others, and that person would tell others. Therefore, the actual cost may be 1/3 of the current cost.

Develop B-end users

In the future, David plans to release additional generators and work hard to expand B-side users. Since most of the traffic previously came from social media, most of the users of formulabot are young people. To this end, David plans to attract older users with decision-making ability to promote more team subscriptions, which currently only account for 2% of subscribers.

David plans to continue working alone , but may consider hiring a full-time employee if he encounters a problem he can’t solve or someone who can complement his skills (such as helping to drive B2B sales).

Advice for Startups

Don’t start from scratch when designing a website. Instead, learn from successful SaaS websites and get inspiration from their designs. This is because they may have done a lot of UX testing to get to where they are today. Entrepreneurs should take advantage of this and learn from several different websites to build something different for their own website.

When David first built the website, it was free, unlimited, and did not require login; then, users were required to log in; then, usage was limited, and finally it became what it is now. When people experience unlimited use, they will become very dependent on it, and then set up payment, the conversion rate is very high. This strategy can also be considered for the free trial period of new products.

David believes in allowing users to experience all the features of the product completely for free (Free Trial model), rather than limiting certain features (Freemium model) – if users don’t try these features first, how will they know if they need them?

For more about David’s story, please visit our Youbute channel at: https://www.youtube.com/watch?v=Nrx-NDjEN7w.