Blog | The data of growth

Last week, I reflected on our goals for growth. I looked back through the history of Mihaly Slocombe and explored the decisions we’ve made as we’ve expanded from two to eight people.

I concluded that we began growing without a systematic reason to do so. We had a little more work than we could handle, so we took on an almost-graduate. We continued to grow because our workload continued to increase. But we’re still growing because we want to buy all the best equipment. Or in architectural terms, we want to do cool things like earn decent salaries, diversify, broaden our expertise, undertake design research, make models, take on leadership roles in the profession, and embrace the opportunities of the 21st Century.

This reflection was a qualitative one, and I remained curious whether we have any hard data that might illuminate the underlying logic of our intuitive employment decisions. I also wondered whether this historical data might hold important lessons for our future.

But first I had to determine which data would be useful.

Gross numbers, like our total active projects or monthly invoice count, would be of little help as they mask the relationship between productivity and income. Instead, I needed data sets that corrected for the number of people in our studio.[1] I decided I would look at the following three sets:

• Number of projects per FTE
• Number of forecast monthly invoices per FTE
• Number of actual monthly invoices per FTE

In past articles, I’ve analysed our income in various ways (here, here and here). This time, I was more interested in productivity. I figured that our monthly income would provide muddy data, confused by the profitability of individual projects. Instead, I hoped the above sets would give a clearer idea of our busy-ness, of how many projects we need to balance in a month.[2]

Here’s what I discovered:

There is a lot of fascinating insight here, let me unpack it:

Averages

• On average, we have 4 projects in the studio per FTE. This average has not changed over time.
• On average, we forecast that we’ll issue 2.8 monthly invoices per FTE.
• Prior to employing Eliza, our forecasts were on average 3.5 monthly invoices per FTE. Since employing Eliza, this average has decreased to 2.0.
• On average, we issue 2.8 monthly invoices per FTE.
• Prior to employing Eliza, on average we issued 3.4 monthly invoices per FTE. Since employing Eliza, this average has decreased to 2.2.

These averages indicate that our workflow has smoothed out since employing Eliza, that is, individual team members work on fewer projects now than they did then. Over this same time period, it’s worth noting that our average monthly income per FTE has gone up by 7%. This means that while we’re working across fewer projects per FTE, we’re (marginally) more profitable.

Spikes

• We employed Eliza and Lauren following substantial spikes in the number of projects, forecast monthly invoices and actual monthly invoices per FTE.
• We spiked at 5.7 projects per FTE prior to employing Eliza, and 5.6 prior to Lauren.
• We spiked at 3.5 forecast monthly invoices per FTE prior to employing Eliza, and 3.3 prior to Lauren.
• We spiked at 5.5 actual monthly invoices per FTE prior to employing Eliza, and 3.3 prior to Lauren.
• For both Eliza and Lauren, the spikes were sustained across successive months.
• We employed Job, Amiee and Leanne following smaller but still noticeable spikes.
• We spiked at 4.1 projects per FTE prior to employing Job, 4.2 prior to Amiee and 4.3 prior to Leanne.
• We spiked at 2.6 forecast monthly invoices per FTE prior to employing Job, 2.2 prior to Amiee and 2.2 prior to Leanne.
• We spiked at 2.1 actual monthly invoices per FTE prior to employing Job, 2.7 prior to Amiee and 2.2 prior to Leanne.

These figures indicate that prior to employing Lauren, we got itchy feet when we were working on more than 3 projects per month per FTE. Since employing Lauren, the spikiness of the graphs has diminished, meaning we now get itchy feet when we’re working on more than 2 projects per month per FTE. I think this is evidence of the normalising effect of a larger team on our workflow.

So what does all this tell me about our future?

For starters, I now have some numbers up my sleeve that tell me what our trigger points have been in the past to employ new staff. It will be useful to fold this analysis into our monthly invoicing cycle as a way of proactively monitoring our productivity. I also now have some revealing ratios between the projects we have on our books and the projects we actively work on in any given month.

Beyond this, I’m not sure if I’ve discovered anything concrete, but I am now more aware of the questions we should be asking: will our recent readiness to employ new staff at a narrower projects:FTE ratio continue to narrow or will it stabilise? Does it make more sense to employ new staff when we’re not busy, briefly busy, or busy for a protracted period? If we decide we don’t want to grow anymore, how many new projects can we afford to take on? And if we do want to grow, what rate of new projects do we need?

As I write this, one of Melbourne’s hottest young construction companies, Project Group, has just gone into liquidation. We worked with them on a small apartment development a few years ago, not that long after they started out. And then very quickly, they were everywhere. I heard from another builder recently that their annual turnover peaked at $140m. But suddenly they were gone, leaving behind around $50m in unpaid debts.

Now, Mihaly Slocombe is in fine financial health thank-you-very-much, but if Project Group was going gangbusters until suddenly it wasn’t, perhaps we need to be paying attention. So perhaps the most important questions that have arisen out of this analysis are: what ratio between projects and staff is sustainable for our business? And in the broadest sense, how can we ensure that our growth retains this sustainability?

This and my last article on growth can be read in conjunction with a couple of Dave Sharp’s recent blog articles, Could architects make more money? and Can architects beat the builders? They explore similar issues and together with mine were published as a loose question and response series on growth, income and market share.

---

1. This type of analysis is known as cohort analysis and compares the internal relativity of related factors. For example, if our gross earnings had grown from $10,000/month four years ago to $50,000/month this year, there’d be high fives all around. But if during that time, we had also grown from 2 FTE to 20 FTE, our earnings per FTE would have shrunk from $5000/month/FTE to $2,500/month/FTE. No high fives. I encountered this concept in Eric Ries’ The Lean Startup, which is essential reading.
2. Clarification: the number of invoices we issue in a month is not always the same as the number of projects we work on. But it does give a good approximation.

Images:

1. Up; author’s own image.
2. Staff analysis; author’s own image.
3. Projects per FTE analysis; author’s own image.
4. Forecast invoices per FTE analysis; author’s own image.
5. Actual invoices per FTE analysis; author’s own image.