In various content I do, the subject of strong and weak assumptions comes up from time to time, which we can also call assumptions. Those with mathematical backgrounds are much, much more sensitive to this issue than those with business and economics backgrounds who embark on the creation of a proto-company, that is, a start-up. Getting assumptions wrong, the “data” that is taken for granted, can nip companies in the bud, and even mature ones: just think of the Signa Group and many other companies that took low interest rates for granted or that bundling dangerous mortgages with other loans would hold up. This is where extreme events come in, if you have listened to some of the podcast episodes. The world rarely adapts to assumptions, which is why monitoring and risk management exist. This applies to statistics, something abstract if you will, from how many cows to keep alive during the winter to the cost of having no grain for the duration of the winter, that is the issue of managing profits and maintaining capital.
I remember that after the outbreak of COVID in Italy, in March 2020, several people were launching statistical models to predict the number of daily cases of new infections. I had two professors who did the same, but in their article they were covering their backs, suggesting not to trust the results, as the data that trained the model had various limitations. They had rightly recognized that the assumptions, on the data, on which they built the model, were not solid. In the case of COVID, these models in the worst case led to a health management that caused more deaths. In the request that inspired this article, the assumptions lead to a different business management and a greater risk of death of a business project.
As a premise of the request that inspired the article, I specify that outbound means contacting the potential customer, inbound means when the potential customer comes to you. TAM means available market, in this case as a number of companies. Sequences, in the digital technical market, means a series of emails sent to a potential customer. ARR means annual recurring revenue. Demo presumably means a trial for a service they offer via web app (business model called SaaS).
Outbound has become the primary revenue driver for our startup, and we need to shape it the right way. We have identified the companies in our TAM and the right contacts to email. We send them highly customized sequences with the goal of generating demos.
Based on the following assumptions, this case study aims to build a model to predict the new monthly ARR generated by the outbound channel.
We need to answer the following questions:
– How many demos can we generate with outbound?
– How much can we expect to close with outbound?
Expected outcome: model in your preferred software: Excel, spreadsheet, Causal, etc.
– Number of demos booked through Outbound per month
– New MRAs closed by Outbound per month
Assumptions
– Outbound channel start date: January 2024
– Total companies in TAM: 100,000
– Email volume capacity per month: 120,000
– Minimum number of emails to be sent per company: 3
– Number of emails per sequence: 4
– Unsubscribe rate: 1% per sequence
Total number of sequences before letting a firm stay: 4
– Cumulative interest rate after these four sequences:
(A sequence is sent to a company every two months)
– Month 0: 2%
– Month 2: 4%
– Month 4: 6%
– Month 6: 8%
other assumptions
– Closing rate: 20% (probability of closing or selling within the month)
– Deal size: $10,000
Definition:
– Interest rate = company booking a demo / Company contacted via email
– Cumulative interest rate: this means that if I send an email to 100 companies in
January, two will book a demo in January, two more in March, two more in May and two more in July – 7 months later, 8 out of 100 companies will have secured a demo.
– Closing rate = Company signing up for service / Company booking a demo
– Sequence = Series of 4 emails to be sent to a company in a given month.
– Email volume capacity = how many emails we can send in a given month.
– Outbound channel start date = the month in which we will contact the first companies for their first sequence.
As a short comment I can say that I find the numbers of the hypotheses extremely optimistic, but maybe they have a marketing department with a budget of 1M dollars per year and they didn’t write it.
We are used to thinking in a linear, additional way. Progressive interest rates can infatuate some novice investors, not the more experienced ones. Because they know that entrepreneurial adventures have ups and downs (non-stationary historical series for friends), and only after having aggregated for several months, assuming that the company still exists, can the interest curve be smoothed into something that resembles a straight line, therefore something progressive.
More truthful assumptions
Having more accurate and truthful assumptions available can cost a lot in terms of researching or acquiring the experience of others. And understandably, those who are starting a business are unlikely to pay someone to tell them they are about to be ruined. And this is also why most startups fail: bad hires, possibly coming from an information gap and an excess of self-confidence.
Statistics can scale back hiring, which may not appeal especially to entrepreneurs with narcissistic traits. It has at least two ways: use the company’s history, even if it is limited (as the case above) and thus would lead to wide ranges, or use public historians. In the second case as a database we have ISTAT, but there are also others. Obviously the paid ones can make more surgical corrections.
But there is an even more fitting solution, which unfortunately startups can hardly afford: market analysis. If there are companies that deal entirely with this, evidently there are still companies out there that want to hurt themselves less or anticipate suffering.
If you have an unverified assumption of your customers’ path to purchase, or the market size of your product or service, excluding e-commerce, we can first discuss it in a free call, and then structure ongoing statistical advice.