# Three Magic Numbers

Every company I’m involved in keeps track of numbers. Daily numbers, weekly numbers, monthly numbers. Ultimately, all the numbers translate into three financial statements – the P&L, Balance Sheet, and Cash Flow Statement. While these numbers are sacrosanct in the accounting and finance professions, they are lagging indicators for most startup companies. Important, but they tell the story of the past, not what is going on right now.

I’ve formed a view that every young company should be obsessed about three magic numbers. Not two, not five, but three. Before I explain what those numbers are, I need to tell a story of how I got to this point.

My brain works better with numbers than graphs, so over the years I’ve conditioned most people I work with to send me numbers on a regular basis. Words are good also, but I love numbers. Early in the life of the company I request numbers daily. Some of this is for me; most of it is to try to help the entrepreneurs build some muscles around understanding the data and how to use it.

Recently, I’ve noticed a cambrian explosion of data among several of the companies I work with. The number of different numbers being tracked daily is massive. When you walk into their office there are screens full of graphs on the wall. Everyone in the company has access to the trends over time across a number of dimensions. These graphs are pretty, the numbers are dynamic, and there are often blinking lights to go along as a bonus.

A few months ago I stood in the middle of the office of a 30 person company and stared at the flat screen TVs hanging from the ceiling showing an array of graphs. I’m sure my mouth was open as I tried to process the data and make sense of it. I knew this particular company well and could reduce the number of different data points to a small set, but I was completely overwhelmed by the visual display. As I systematically looked at each of the graphs, I realized very few of them mattered much, nor where they particularly helpful in understanding what was going on in the business.

At the moment I realized these were no longer magic numbers. Instead, I was looking at wallpaper. Data porn. The entrepreneurial aeron chair equivalent of 2012. Pretty, but a bad allocation of resources. The 30 people in the room might be looking at the graphs. They might be looking at one of the graphs. But they probably weren’t seeing anything.

This particular company runs off of three numbers. Daily active users (DAU). Live publishers. Trial publishers. That’s it for now. In the future, there will be a daily transaction metric (Daily transaction revenue) that replaces trial clients. But that’s probably a quarter or two away.

I then started thinking about each company I’m on the board of. This rule of three applies. For many of the companies, DAU is one of the numbers. In others it’s daily orders. Or daily revenue. Or daily activations. Or total publishers. Or new publishers. But in every case I could reduce it to three numbers that I felt were the most important to pay attention to.

The absolute number is what matters. The trend is driven by day over day changes. If during the week (assume the week starts on Sunday) the numbers are 47, 67, 69, 72, 174, 80, 53 this prompts the question “what happened on Thursday to drive the number to  174?” If the next week the numbers are 53, 75, 214, 83, 80, 73, 45 this prompts two questions: “what caused the spike on Tuesday” and “why is the week over week trend downward?” Clearly there is seasonality within the week and there is a new high, but the overall trend going into the weekend is negative.

My brain can focus intensely on three variables like this in a business. Once I add a fourth, I have trouble figuring out the relationship between them. This doesn’t mean that the leadership and functional managers shouldn’t track and analyze the detailed data. They should. But they should realize that when they show this to everyone in the company, no one knows what to care about.

Instead, my new approach is to focus on three numbers. These three numbers should reflect “what’s going on right now in the business” and the trend of the numbers should be a predictor of what’s going on. As I think about the companies I’m involved in, I can define these three numbers in 60 seconds – they are almost always painfully obvious. Sometimes I do end up with four and have to make a choice, but I rarely end up with five.

The technology for displaying these three numbers is remarkably simple. They make this thing called a whiteboard that you can write them on. An email can go out to everyone in the company with the three numbers. That’s it.

What are your three magic numbers?

• While we aren’t necessarily there yet, I love zeroing in on what really matters. I feel that our three magic numbers might be 1) # of Users/Dealers, 2) # of Active Applications/Bids, and 3) % of Accepted Bids/Appointments. This of course complimented with accounting porn (P&L, Balance Sheet, and Cash Flow). These magic three numbers will tell a story of how well we are placed in the market and our real-time traction. More importantly, will hint as to when a pivot might be necessary. Thanks for the thoughts Brad! (re: Bidzuku http://bidzuku.com)

• 1.) New subscribers   2.) total members   3.) incoming connections

• For Blogmutt I have two of them at the top of my admin panel: Number of Active Subscribers (meaning they are currently paying), an Number of Active Writers (last 60 days.) The third is the number of customers’ queues with no posts in them. We want that number to be zero, so we need to keep the second number somewhere within a range of the first number. And we want the first number to grow every week.

• Bingo!

• JMoran

I appreciate this article – Will certainly pass this along in client engagements.

• Anonymous

Ok, funny that you pick three.  My boss once said that everything can be boiled down to three easy to remember points.  Someone cracked that kids took more than three points, at which someone said “Sit down, shut up, and be quiet.”  Really a watchword in our group.

I once inherited a group that had “operating metrics” on a twelve tab spreadsheet.  It was a big group and did a lot of revenue, so I plunged in.  Turns out we needed three things: engagement accepted/rejected ratio, utilization, %target achieved.  Also turns out we were spending 7% of our total time in overhead tracking stuff we didn’t really need.

The rule of three indeed.

-XC

• Great example of getting lost in the endless weeds of a spreadsheet.

• Nobody remembers more than 3 things.  Make them count.  I tell my GPs to focus their message down to 3 hooks for fundraising, annual meetings, or company updates.  3.  Only 3.

• John Whitcomb

Brad, Three is the lovliest number!  I think that goes for all business analytics and strategy.  Somehow human aptitude and human attention span are both limited to three underlying ideas per subject.  Mention a forth and the three before are reduced in impact – and the human mind does not so easily wrap around the concept of storyline.  I beleive this applies to three areas of business:  One is as you say to the business metrics.  Second is to marketing that should be segregated into three features, three benefits, three main customer groups, three product groups or versions… and so on.  The third area is business issues, what are the three biggest issues, three best solutions, three perspectives that lead to better understanding understanding. Everything does not matter – Occam’s Razor and the law of parsimony, economy or succinctness. Thanks for sharing your learning.  Have a great day – I hope to see you soon –  http://en.wikipedia.org/wiki/Occam’s_razor

• I don’t remember his three criteria, but Scott McNealy, former CEO of Sun Microsystems, said he could monitor all of Sun’s business based on his three variables.

I also worked with a large international pizza chain, and they had a  similar “scoreboard”. With one touch on a keyboard in a store, you could see what they considered to be the vital operating stats of the company (using numbers, not graphs).

• Great post Brad. It is easy for all of us to get lost in data! That which is measured is improved….everyone needs to get this right!

• Totally agree with the post.   I would modify the sentiment just a little however.

Lijit (as you know) was very focused on publishers and page views over most of our history.  This served the company well i believe because it gave everyone a common and understandable number to drive that had fundamental results for the business.   As the company got larger and the relationship between numbers, ie (page views, ad calls, fill rate, CPM) became more clear it started to become advantageous to tailor the three numbers to different groups within the company.   Now publisher recruiters drive ad calls.  Business managers drive revenue through the site quality, fill rate, and product mix.   Today the business drives to three numbers for the CEO but some of these other numbers are much more actionable to other parts of the company. I think its important that each group know its three numbers but the company may have more than three, at least as it matures.

• Right on – as you scale up, you start to have different sets of three numbers. You describe it beautifully.

• Ha ha! Money phrase: data porn. Long live the magic marker. 🙂

• Data porn. +1

• i used to

• Definitely agree that excessive ‘data porn’ (great term) actually harms productivity, but I just wanted to share an example of a  beautiful status board that really does help.

Panic (makers of Coda, Transmit, etc) use this to share the most important information for their business; unsurprisingly, it boils down to around three things.

http://www.panic.com/blog/2010/03/the-panic-status-board/

I know Square also has a similar setup that allows them to live-track how much money they’re processing and the revenues being generated.

• Beautiful!

• As of how to best communicate the numbers – Gdocs is the best way to keep the team up to date and on the same page. An always open browser tab for Gdocs works like a communication tool in itself – everyone sees when a doc has been updated (no need for email notifications).

• i used to race things.  bikes and skis.  one thing i noticed again and again over the years: by and large the guys with heart-rate monitors, computers, and numbers never trained hard enough.  they sucked.  they always wilted when it got really hard.  i see the same thing with engineers and tech people.  they measure their inbox.  their stories.  their revenue.  all if it is a distraction.  if can be an error to measure things only objectively.

in exercise phisiology they know that perceived exersion is vastly more accurate than any other measurement the lab guys can come up with.  but still, the weak hunt for numbers.  the super strong go out and try to suffer ‘more’.

taking it back to tech: why don’t we consider customer perceived quality, developer perceived quality, designer percieved quality, manager percieved as important numbers?

i think i know the answer: because simple numbers are a great excuse to say “see, i tried hard enough, i can go home now.”

i love the kiss idea.  but i think the field as a whole misses the boat in it’s continual search for purely objective numbers.

• JamesHRH

Your ‘hero athlete’ approach does not scale.

It has awesome positive aspects, but it does not work in a group setting – everyone can push themselves in a heroic way, but it is the combined effort that produces progress (and heroic efforts often go against each other).

What Brad is saying is – cut the BS. 3 things matter: measure them, think about what the measuring tells you, act.

That’s the team version of your philosohy.

• This is a good post and makes a lot of sense.  During the process of selecting the 3 or maybe 4 metrics that will be used requires management to think about what really drives the business, and ultimately how they make money.  Being able to understand, track and measure the core of the business is important to make accurate business decisions.  Choosing the metrics is an important step in any industry.

I enjoy the stock market and investing, and this same principal applies when researching a company and determining if it is a good value.  I read the 10 Q’s and K’s and try to understand the essence of the business and how the business makes money.  From there, you can determine which ratio’s (metrics) that are needed to value the business and determine how the business is performing.  This is typically only only 3 or 4 ratios using the financial statements.

• Editing is an insanely hard life skill to learn, especially where we’re conditioned to keep consuming more. It’s usually only after gorging that we learn to eat the right things.

• You are spot on. I would add that if time series numbers are the key metrics (which they usually are with startups) knowing a little about data analysis concepts called arima and stationarity will help make better use of the data and avoid key errors in its interpretation.  This area of analysis earned Robert Engle a noble prize! As to the specific metric,      this is where the fun begins, is it easily captured, are there industry comps, does the metric scale linearly…etc. test and iterate!

• plumbing-heating-denver.com

thought your insight was great and enjoyed the blog

• Whats interesting is that the human brain tends to remember things in three. especially numbers. Way to modify your thinking to improve the assimilation and understanding (and remembering!) of distinct objects. (Oh, and [email protected]:disqus said.)

• Whats interesting is that the human brain tends to remember things in three. especially numbers. Way to modify your thinking to improve the assimilation and understanding (and remembering!) of distinct objects. (Oh, and [email protected]:disqus said.)

• No doubt plays right [email protected]:disqus ‘s infamous top of mind, which only comes in threes.

• Indeed!

• Regina Spaikerhen

Right, absolutely! My university teacher of Phylosophy used to say us the same thing about threes:)

• This could be a slippery slope, but do you think each business unit should have its own three metrics? Such as operations focusing on uptime / back-end metrics, while biz dev has its three, and so on…  Or rather should the company have three, period?

• I think there should be three high order metrics that everyone in the company sees. I think each business unit / department can then have a different set of three. But no one person should be processing more than the “top level metrics” and the core group metrics.

• I like this.  I’m in (re)start-up mode, and that has forced me to reevaluate many of the things I did wrong a few years ago.

I’m going to spend some time today and thoughtfully decide what my top 3 company-wide numbers are.  These numbers are not immediately obvious, for us, the way it is with many of the Web 2.0 folks.

We are a small team, just 8 of us.  I’ll ask each person what their – individual – 3 numbers are as well.

Thanks.  Take care., mjl

• CarlB

An mature individual’s brain can process up to seven things effectively (more during developmental years) w/o gimmicks (mnemonic anchors/triggers).  And the absolute guru for all “graphical representation of statistical data,” Edward Tufte, would argue they’re not presenting properly if a state of MEGO (My Eyes Glaze Over) or mouth breathing occurs, as you describe.  But the most effective teams, companies, _______ (fill in blank), distill all various work-streams to only three current and relevant objectives *at that point in time*.  2005/06 T-Mobile had six major initiatives on a new prod revenue growth plan (I was on one…UMA PMO for quality), however meeting overload and hour-long overlap would occur if you couldn’t distill the updates, issues and remediation adjustments to just 3 things at a time.  Under those rules, the mantra was finish and send everyone back to where they work best (even collaboratively)…THEIR SPACE!  Violate the rule and you’d waste time you couldnt’ get back on the gant chart, or in your life.  To this day I’ll catch myself humming in Op’s meetings, Finance and BOD mtgs, etc,  “Schoolhouse Rock…3 It’s a Magic Number!”  http://youtu.be/aU4pyiB-kq0

• JamesHRH

Carl – I think you are misquoting the 7 rule. People cannot “process” 7 things. They can on average, remember 7 items as a list.

Toothpaste: Crest, Colgate,…………..

• Anonymous

Awesome post Brad.  I blogged about a similar topic that I called the 3 Keys a few years back with regards to what I call “Order of Magnitude CEOs/Founders”. Of course I included 3 Founder/CEO examples (Amazon, Southwest Airlines and Nu-Kitchen )
Here is the post in case you are interested (it’s short):What Are Your Three Keys?http://bit.ly/djpdah

• oneafrikan

We’re using Geckoboard (http://www.geckoboard.com/ and paying for it) and find it really helpful.
We’re an ecommerce business, so numbers are:

1) Conversion Rate
2) Average Order Value

All 3 have acceptable variation, and all 3 have targets we’re working towards.

We’ve got a bunch of dashboards for different silos, but those are the key numbers which drive our economics.

• Anonymous

We use three sets of metrics with three sub categories for each…

– Total Impressions (Daily, Weekly, Monthly)
– Revenue (Total Rev, eCPM, Average Monthly Spend Per Client)
– # of clients (Recurring, Trial, Seasonal)

• We focus on two things – growth and revenue.

Growth – (net subs per day, % of new subs for each product)
Revenue – (deal RPM, daily revenue/profit)

• Hey Brad – how did you come to the conclusion that that company had poorly allocated its resources to all those graphs etc?

Reason I ask… did they do a bunch of extra work that probably won’t pay off and distracts or do they build metrics into every part of the system as a general product philosophy in the hope of seeing signal pop out of the noise? maybe they just have bad noise filtering ..and hence the data porn 🙂

• It’s a combination – too much energy going into instrumenting and collecting the wrong data, but more importantly just a bunch of data noise that no one is now paying attention to (or at least not the right stuff.)

• working on a Dashboard that has flow. Current dashboards put the numbers and charts in boxes and makes you think in terms of boxes. There is no flow of any kind. watch out InfoCaptor Web app coming soon http://www.infocaptor.com

• The idea of tracking three key numbers is great.  But those are just the tip of the iceberg, and as a VC that might be all you are interested in.  But I am going to want to be able to drill down deeper to see what is driving those three magic numbers, so I can have some chance of improving on them.  And for that I think you need something a bit more sophisticated than a whiteboard.

• I definitely agree with that. But do you want every single person in your company drilling down? I doubt it. You might want them to drill down on a particular number, or set, but not everything. And – if you have them drilling down on everything, I can almost guarantee that they will be confused and not be able to see the forest for the trees.

• Who needs it is highly specific to the business.  My experience is in the trading field where everyone on the desk definitely needs it and wants it (with excruciating detail), but management needs to exercise control over who sees what and when.

But my point is that if even one person needs to drill down on these key metrics you will need to invest in some system for collecting data, aggregating and studying it.  As a firm, if all you have are three numbers then I think you are flying blind.

But I agree that as a VC, you probably cannot handle more than 3 numbers per company.  If you have 30 companies that is still a lot of data.  Also, as a ceo identifying and fixating on those three numbers is essential.

• I think you are misinterpreting what I’m saying. Every company I’m involved in is heavily instrumented. The amount of data being generated is spectacular. Many of the companies integrate this data directly into their business (e.g. the machines are running the machines, not the people.) However, when a broad cross-section of this data, in great detail, is shown to everyone in the company, it’s pretty easy for folks to get confused about what matters.

• Derek Scruggs

This makes sense at the executive and board level, but remember departments need metrics to work against, too. So while you may not care about, say # daily of commits, it’s a useful metric for a company trying to be agile and the kind of thing that might be charted and reviewed at the quarterly offsite.

• True, although I think the entire company can get aligned around three numbers. For example, if devs are focused on daily commits, it’s pretty easy to see how that generates a particular focus that might not actually move things forward in any material way.

• Pingback: Three Magic Numbers « Elia Insider()

• DanielleDB

Great post.  I would just add that some sort of moving average should be applied to the numbers and that choosing the time length of the MA (1 hr, 1 week, 3 months, etc) is as important as the numbers themselves.  When your numbers break through the MA, up or down, that’s when you need to pay attention.

• Pingback: Three Magic Numbers – SchemaByte()

• Pingback: Lessons for Data Driven Businesses « Mr Gee()

• Pingback: “数据驱动型公司”如何巧用数据做决策？ | 云动态()

• Pingback: “数据驱动型公司”如何巧用数据做决策？()

• Pingback: “数据驱动型公司”如何巧用数据做决策？ - 阅趣-阅读的乐趣()

• Pingback: xbox 360()

• Pingback: mike geary truth about abs()

• Pingback: qualitywebhostings.com()

• Pingback: bumperstickerquotes.org()

• Pingback: win a free ipad()

• Pingback: best led tv brand()

• Pingback: free xbox()

• Pingback: dual screen portable dvd player()

• Pingback: truth about abs scam()

• Pingback: Stefany Lonneman()

• Pingback: Vaughn Romera()

• Pingback: win a free ipad()

• Pingback: best 55 led tv()

• Pingback: portable dvd player reviews()

• Pingback: the truth about six pack abs()

• Pingback: Thalia Vogds()

• Pingback: Jerrell Honokaupu()

• Pingback: Odilia Tochterman()

• Pingback: Aretha Knightly()

• Pingback: Cristobal Baughman()

• Pingback: Irvin Canela()

• Pingback: Willis Morre()

• Pingback: Roy Hequembourg()

• Pingback: Enrique Tarascio()

• Pingback: Stevie Saleeby()

• Pingback: Emely Cardno()

• Pingback: Leda Tepe()

• Pingback: Kareem Henton()

• Pingback: Cecile Gerson()

• Pingback: Jena Donchatz()

• Pingback: Marion Farley()

• Pingback: Noel Demman()

• Pingback: Elene Luing()

• Pingback: Jacquie Matheny()

• Pingback: Geoffrey Giorgi()

• Pingback: Paul Santio()

• Pingback: Pat Fraser()

• Pingback: Ignacia Toolsiram()

• Pingback: Portia Vanderen()

• Pingback: Shelby Rasbury()

• Pingback: Rita Linzan()

• Pingback: Malika Spiers()

• Pingback: Lilla Merk()

• Pingback: Rocky Gasquet()

• Pingback: Libby Zboral()

• Pingback: Eugenie Bennick()

• Pingback: Darell Leischner()

• Pingback: Martine Mazzarella()

• Pingback: Raymond Shireman()

• Pingback: Alayna Gelle()

• Pingback: Paris Kinkle()

• Pingback: Wilburn Niglio()

• Pingback: Nga Liszewski()

• Pingback: Alica Graci()

• Pingback: Sparkle Plesant()

• Pingback: Otis Rathai()

• Pingback: Sandy Besong()

• Pingback: Georgeanna Michie()

• Pingback: Mohamed Winter()

• Pingback: Cortez Bedingfield()

• Pingback: Steve Nivar()

• Pingback: Alex Jerrett()

• Pingback: walks In()

• Pingback: iOs()

• Pingback: Jame Tefera()

• Pingback: Devon Krenn()

• Pingback: Aimee Bridenstine()

• Pingback: Rana Scheidler()

• Pingback: Meg Nemard()

• Pingback: Numbers Jacckson()

• Pingback: Nohemi Than()

• Pingback: Glen Mallo()

• Pingback: Cecil Shau()

• Pingback: Farrah Blazek()

• Pingback: Nia Nowell()

• Pingback: Christinia Hoist()

• Pingback: Val Zody()

• Pingback: Terra Florez()

• Pingback: Vallie Morrone()

• Pingback: Jimmie Broz()

• Pingback: Kendrick Kutsch()

• Pingback: Hollis Liska()

• Pingback: Emanuel Polakowski()

• Pingback: Reggie Hallmon()

• Pingback: Lorean Tkach()

• Pingback: Oscar Yannantuono()

• Pingback: Donn Sudweeks()

• Pingback: Lilia Chalaban()

• Pingback: Kurt Fedder()

• Pingback: Karole Eilderts()

• Pingback: Octavia Degunya()

• Pingback: Elroy Agbisit()

• Pingback: Ophelia Salvesen()

• Pingback: Garrett Martiniz()

• Pingback: Takisha Lamson()

• Pingback: Drew Cherrette()

• Pingback: Alonso Connoly()

• Pingback: Ashleigh Limbaugh()

• Pingback: Earlean Bargo()

• Pingback: Yvette Steuck()

• Pingback: Herb Chinick()

• Pingback: Pansy Ayaia()

• Pingback: Al Robberson()

• Pingback: Troy Heafey()

• Pingback: Marc Kjellsen()

• Pingback: Elwanda Molineaux()

• Pingback: Autumn Coletta()

• Pingback: Toni Fiereck()

• Pingback: Bernetta Grapes()

• Pingback: Darren Bann()

• Pingback: Lara Hass()

• Pingback: Taneka Dipiano()

• Pingback: Sydney Benejan()

• Pingback: Dalila Dyal()

• Pingback: Carol Copping()

• Pingback: Andre Deller()

• Pingback: Don Ryks()

• Pingback: Dovie Ahlheim()

• Pingback: Olin Obanner()

• Pingback: Julian Melzer()

• Pingback: Nathanial Camarena()

• Pingback: Elke Elnicki()

• Pingback: Sammy Hutti()

• Pingback: Harold Yeargin()

• Pingback: Keven Natalizio()

• Pingback: Wesley Marye()

• Pingback: Christopher Naas()

• Pingback: Bruce Ciancio()

• Pingback: Bert Yett()

• Pingback: Lien Petosa()

• Pingback: Hilario Gnagey()

• Pingback: Noel Shy()

• Pingback: Allen Corbin()

• Pingback: Wyatt Morneau()

• Pingback: Lakenya Nobbe()

• Pingback: Ethelene Gates()

• Pingback: Colin Smyntek()

• Pingback: Tonja Magness()

• Pingback: Darlene Slovinski()

• Pingback: Calvin Koslow()

• Pingback: Cleveland Secundo()

• Pingback: Janeen Remedies()

• Pingback: Rupert Simison()

• Pingback: Maragaret Challa()

• Pingback: Jovan Staberg()

• Pingback: Parker Isenberg()

• Pingback: Carlton Kloeppel()

• Pingback: Marvin Filbert()

• Pingback: Tracey Campanella()

• Pingback: Elden Victoria()

• Pingback: Jeffrey Doporto()

• Pingback: Ernest Kuboushek()

• Pingback: Christeen Salazan()

• Pingback: Dennis Sylve()

• Pingback: Susann Dartez()

• Pingback: Kieth Krieger()

• Pingback: Noble Ancheta()

• Pingback: Chance Oldroyd()

• Pingback: Junko Getschman()

• Pingback: Arlen Inabnit()

• Pingback: Nannette Nyswonger()

• Pingback: Edwin Eggeman()

• Pingback: Tessa Burgh()

• Pingback: Clarence Ruka()