Λόγω δουλειάς, το 2015 βρέθηκα να μοιράζω το χρόνο μου ανάμεσα στη Βοστώνη, το Λονδίνο και την Αθήνα. Αυτό που μου έκανε μεγαλύτερη εντύπωση είναι πόσο διαφορετικά είναι τα πράγματα που πιάνω τον εαυτό μου να διαβάζει, και να συζητά σε κάθε μία από αυτές τις τρεις παρόμοιες πόλεις του δυτικού κόσμου.
Στην Αμερική το τέλος της βιομηχανικής επανάστασης θεωρείται δεδομένο. Οι μεγαλύτερες εταιρίες της χώρας είναι τεχνολογικές. Ηλεκτρικά αυτοκίνητα κυκλοφορούν ήδη στους δρόμους και το ερώτημα είναι τι θα γίνει σε λίγα χρόνια που τα περισσότερα θα οδηγούνται από software. Η αγορά εργασίας αλλάζει λόγω των υπηρεσιών on-demand και της εργασίας από απόσταση. Υπάρχει μεγάλος δημόσιος και ιδιωτικός διάλογος για την μεταμόρφωση των μέσων ενημέρωσης, τη νομοθεσία για εμπορική λειτουργία drones, την απεξάρτηση από το πετρέλαιο και τις επιπτώσεις της τεχνητής νοημοσύνης για την αγορά εργασίας. Αυτά και άλλα πολλά είναι περίπλοκα θέματα που απασχολούν τους πολιτικούς, τους σχολιαστές και τα μέσα ενημέρωσης γιατί θα καθορίσουν το μέλλον μας.
Στο Λονδίνο πολλοί ανησυχούν αν η Ευρώπη μένει πίσω από την ψηφιακή επανάσταση. Οι τρεις μεγαλύτερες εταιρίες internet της Ευρώπης αξίζουν μόλις $25 δις ενώ στις ΗΠΑ αξίζουν $750 δις και οι αντίστοιχες Κινεζικες $500 δις. Την ίδια στιγμή η βιομηχανική παραγωγή του 20ου αιώνα έχει ήδη αποδημήσει προς τον αναπτυσσόμενο κόσμο και οι παραγωγικοί πυλώνες της Ευρώπης ανήκουν στην οικονομία του χτες. Δημόσιοι και ιδιωτικοί φορείς αναζητούν τις δομές και τα κίνητρα που θα δημιουργήσουν Ευρωπαϊκές εταιρίες και προϊόντα παρόμοιας εμβέλειας με την Silicon Valley.
Πίσω στην Αθήνα, η επικαιρότητα και ο ιδιωτικός διάλογος κινουνται σχεδόν αποκλειστικά γύρω από μέτρα, Eurogroup, εισφορές, εκλογές και μικροπολιτικές ανοησίες τοπικού ενδιαφέροντος. Όχι πως δεν υπάρχουν τέτοιες συζητήσεις και αλλού. Αλλά στην Ελλάδα του 2015 φαίνεται να έχουμε ξεχάσει οτιδήποτε άλλο. H κοινωνία μας έχει μουδιάσει, όχι μόνο οικονομικά αλλά και ψυχολογικά. Η καθημερινότητα και η επικαιρότητά μας έχει μετατραπεί σε ένα deja vu.
Νομίζω το μεγαλύτερο κακό που έφερε η οικονομική κρίση στη χώρα μας είναι ότι μας ματατρέπει σε μία κοινωνία που ξέχασε, και ίσως φοβάται και λίγο, να ασχοληθεί με το μέλλον. Τη στιγμή που άλλοι λαοί ετοιμάζονται για τις δραματικές αλλαγές που φέρνει η τεχνολογική επανάσταση στην κοινωνία τους, εμείς προσπαθούμε – μάταια – να διατηρήσουμε ένα οικονομικό και κοινωνικό μοντέλο που χρεωκόπησε. Από την εκπάιδευση ώς το ασφαλιστικό, αντί να σκεφτόμαστε πως θα προσαρμοστούν σε μια νέα πραγματικότητα που είναι αναπόφευκτη, συζητάμε, μαλώνουμε, διχαζόμαστε και εξαθλιωνόμαστε για να κρατήσουμε το παρελθόν ζωντανό. Για ένα ακόμα μήνα, μία ακόμα δόση. Τίποτα δεν έχει χρονικό ορίζοντα πάνω από λίγους μήνες, τα μάτια μας είναι στραμμένα στο παρελθόν.
Για να το πω πιο απλά, ζούμε στον κόσμο μας. Και δεν είναι ο κόσμος στον οποίο θα πρέπει να ζήσουμε εμείς και τα παιδιά μας σε δέκα χρόνια από τώρα.
Κάποιοι θα πουν είμαστε μια μικρή και φτωχή χώρα, δεν θα φτιάξουμε εμείς Google, Facebook και Tesla. Δεν συμφωνώ με μία τόσο απαισιόδοξη ανάγνωση του μέλλοντος. Η ψηφιακή επάνάσταση είναι τεράστια ευκαιρία για μία μικρή χώρα με μορφωμένους νέους. Δεν χρειάζονται μεγάλες κρατικές επενδύσεις, φυσικοί πόροι, βαριά βιομηχανία και μεγάλες εταιρικές δομές για να φτιάξει κανείς software. Χώρες όπως η Εστονία, (ξέρατε ότι εκέι φτιάχτηκε το Skype;) έκαναν άλματα ανάπτυξης βασισμένες σε μικρές, έξυπνες τεχνολογικές εταιρίες που μεγάλωσαν και έγιναν καθημερινό εργαλείο για την ανθρωπότητα.
Εμείς στη Workable ξεκινήσαμε με δύο φορητούς υπολογιστές, ένα τραπεζι κουζίνας από το ΙΚΕΑ και τις γνώσεις μας. Τρία χρόνια αργότερα, 3.500 επιχειρήσεις στην Αμερική και την Ευρώπη βασίζονται στο προϊόν μας για να βρούν προσωπικό. Μέχρι το τέλος του 2016 θα έχουμε δημιουργήσει 100 θέσεις εργασίας στην Αθήνα. Έτσι ξεκίνησε και το Taxibeat που συμμετέχει στην παγκόσμια μεταμορφωση των αστικών μεταφορών και έχει βελτιώσει τη ζωή όλων μας. Η Transifex που κάνει αυτόματες μεταφράσεις για χιλιάδες επιχειρήσεις παγκοσμίως. Εκατοντάδες τεχνολογικές startup σαν αυτές δημιουργήθηκαν στην Ελλάδα τα τελευταία πέντε χρόνια. Χιλιάδες συμπολίτες μας δουλεύουν στον ευρύτερο κλάδο. Μόνο το 2015 επενδύθηκαν πάνω από $50εκ. σε Ελληνικές startup, χρήματα που ώς επί το πλέιστον ήρθαν από ξένους επενδυτές.
Θα μας σώσουν μερικές χιλιάδες θέσεις εργασίας όταν η ανεργία μετριέται σε εκατομμύρια; Σίγουρα όχι άμεσα, όμως έχουν έναν σημαντικό ρόλο να παίξουν. Αν περάσετε μια βόλτα από τα γραφεία τους δεν θα ακούσετε συζητήσεις για μέτρα, δόσεις και συντάξεις. Θα ακούσετε για το μέλλον του κόσμου και τι μπορούμε εμείς να φτιάξουμε σήμερα, εδώ, για την οικονομία του αύριο. Αυτό που προσδοκούν δεν είναι κάποια ελεημοσύνη από το Eurogroup αλλά να φτιαχτούν πέντε η δέκα μεγάλα τεχνολογικά προϊόντα παγκοσμίου εμβέλειας στην Ευρώπη, και ίσως ένα από αυτά να είναι Ελληνικό. Οι εργαζόμενοί τους, νέοι και επιστήμονες κατά κανόνα, είναι διατεθειμένοι και μπορούν να ανταγωνιστούν τους αντίστοιχους Αμερικανους και Ευρωπάιους. Δεν τους χρωστούν τίποτα και κερδίζουν το σεβασμό – και τα χρήματά τους – με το μυαλό, την εργατικότητά και το τελικό αποτέλεσμα της δουλειάς τους.
Οι startups είναι ένα μίκρό κομμάτι τις οικονομίας μας. Είναι όμως αυτό που δεν φοβάται να κοιτάξει το μέλλον και επιμένει να διεκδικήσει μια καλύτερη θέση σε αυτό για τον τόπο μας. Το 2016 ανυπομονώ να δω τι θα πετύχουν και πόσοι συμπολίτες μας θα επηρεαστούν από την αισοδοξία τους.
[Το κείμενο δημοσιεύτηκε αρχικά στην Καθημερινή τον Ιανουάριο του 2016]
Όταν έρθει η ανάπτυξη, με τι θα μοιάζει; Το ερώτημα δεν είναι περιπαικτικό αλλά ουσιαστικό. Υπάρχουν πολλοί τρόποι να μεγαλώσει μια οικονομία. Aν θέλουμε να ξέρουμε τι είδους ανάπτυξη φαντάζεται η κυβέρνησή μας αρκεί να κοιτάξουμε τι πολιτικές εφαρμόζει σε σχέση με τις επιχειρήσεις.
Έκανα σήμερα το λογαριασμό των νέων μέτρων για τους εργαζόμενούς μας. Είμαστε μία εξαγωγική επιχείρηση υψηλής τεχνολογίας. Οι άνθρωποί μας είναι πτυχιούχοι και υψηλά αμοιβόμενοι. Οι καθαρές αποδοχές των περισσοτέρων θα μειωθούν από 15% εώς 25%, ποσό που δεν αντισταθμίζεται εύκολα με αυξήσεις – ειδικά όταν τα μισά πάνε στο κράτος. Αυτό υποβαθμίζει την ικανότητά μας να προσελκύσουμε στελέχη υψηλών δεξιοτήτων που απαιτούν μεγάλα εισοδήματα (π.χ. άνω των €50.000) ή την προοπτική να εξελιχθούν μισθολογικά στο μέλλον.
Στο δυτικό κόσμο οι εταιρίες τεχνολογίας δίνουν μάχη για εξειδικευμένο προσωπικό κι εμείς πάμε με τα χέρια δεμένα πίσω από την πλάτη. Έχουμε χιλιάδες μορφωμένους νέους σε μια χώρα με χαμηλό κόστος ζωής αλλά οι φόροι και εργοδοτικές εισφορές εξαλείπτουν το οποιοδήποτε πλεονέκτημα θα είχαμε απέναντι στην ξένη επιχείρηση. Το κράτος λέει στον επιχειρηματία “εδώ θα έχεις μόνο κατώτερα στελέχη, τις καλές δουλειές να τις φτιάξεις σε άλλη χώρα”. Στον εργαζόμενο λέει “αν έχεις ικανότητες διεθνούς επιπέδου, πήγαινε να βρεις δουλειά αλλού.” Και να μην μεταναστεύσουν, τι κίνητρο μπορεί να έχουν να δουλέψουν σε υγιείς επιχειρήσεις που δεν επιτρέπουν τη φοροδιαφυγή; Πως θα αποταμιεύσουν για να επενδύσουν σε μελλοντικές επιχειρήσεις ή στην επιμόρφωση της επόμενης γενιάς;
Ξέρω, δύσκολα θα συμπονέσει κανείς την “πλουτοκρατία” των 50 και 60 χιλιάδων ευρώ που θα πληρώσουν περισσότερο φόρο. Για να το σκεφτούμε όμως λίγο αυτό. Πόσο λογικό είναι; Ονειρευόμαστε δηλαδή μια οικονομία που ευνοεί μόνο την χαμηλόμισθη και ανειδίκευτη εργασία;
Μεγάλες επιχειρήσεις χρειαζόμαστε και για να έχουμε διεθνώς εμπορεύσιμα προϊόντα που φέρνουν έσοδα από εξαγωγές. Αυτές θα εκπαιδεύσουν στελέχη, θα φέρουν κεφάλαια ξένων επενδυτών, θα χρηματοδοτήσουν έρευνα και θα δημιουργήσουν δουλειές για μία μεσαία τάξη μορφωμένων πολιτών δυτικού κράτους. Δουλειές με μεγαλύτερες αμοιβές και προοπτικές εκπαίδευσης και εξέλιξης.
Για να φτιαχτούν τέτοιες επιχειρήσεις είναι απαραίτητο να μπορούν να προσλάβουν στελέχη υψηλών δεξιοτήτων. Αυτοί θα μας ξελασπώσουν, αυτοί θα παράγουν πλούτο και θα πληρώσουν τις συντάξεις των γονιών μας τις επόμενες δεκαετίες. Αυτοί θα ιδρύσουν τις μεγάλες επιχειρήσεις του μέλλοντός μας. Γι’αυτούς θα έρθουν πολυεθνικές εταιρίες από το εξωτερικό να στήσουν διοικητικές μονάδες στη χώρα μας.
Η παλαβή εχθρότητα προς τους “υψηλά αμοιβόμενους” εξωθεί στην παραοικονομία ή στην μετανάστευση το πιο πολύτιμο ανθρώπινο κεφάλαιο της κοινωνίας μας. Δεν ξέρω αν θα έρθει η πολυπόθητη ανάπτυξη. Αν θέλουμε όμως να μοιάζει με δυτικό τεχνολογικό καπιταλισμό, με μεσαία τάξη που έχει ευρωπαικά πρότυπα διαβίωσης και πολιτικής συνείδησης, αυτό δεν θα γίνει με τις πολιτικές που ακολουθούμε σήμερα.
Somewhere between prophecy and ignorance there’s a sensible spreadsheet.Here’s how to build it.
This is a guide on how to build an 18–24 month financial plan for an early stage SaaS company with product market fit, some revenues, growing staff and operations and a financing round in the works. Most of the ideas below will apply to non-SaaS startups, too.
I’ve spent some time doing this for myself and more time helping others and answering questions. There’s quite a few best-practice models out there and templates for SaaS specifically or more generic ones for seed stage companies.
All such templates have a common structure: On the leftmost column you have a P&L-like listing of cost items and revenue-driving parameters, and then you have a column for each of the following 18 months or more.
The exact layout is not so important. Just pick one that you like and use it. What I want to focus on is the thought process: how to analyse and understand what drives your business and how to make sensible estimates for the future.
Assumptions and Models
The most useful part of a financial plan is the parameters and models used to calculate the time series for each line. This is where we can see the business logic and underlying assumptions that govern the entire budget.It’s the difference between a plan and a bunch of numbers.
I use parameters and spreadsheet formulas to model the behaviour of different items in the budget over time. Some items are incremented by an absolute amount or a growth percentage every month while in other cases I want this percentage to be gradually decreasing as time goes by. In some cases a budget item is a function of another item in the plan, such as overheads per headcount or infrastructure costs per customer served.
This allows you to easily tweak the parameters when your assumptions change. (e.g. “what if we added a costly feature that would increase our server costs per active user?”)
A little bit of research or actual historical data will get you a set of sensible assumptions. You won’t nail down everything to the decimal, but you can get fairly good estimates for almost all cost items. We will look at how to research and model each kind of item in the next section.
What matters the most is that you think hard about each item in your list and try to understand what shapes its behaviour. Very few things stay flat, especially in a startup that’s supposed to grow 10x within a year or two. Even if your plan is not totally accurate and some things are unknowns, understanding the nature and dynamics of each item budget is the only way to really know your business.
Perhaps the biggest cost of an early-stage startup, so it’s worth being as accurate as possible with it. Make a line for each role and list the total cost-to-company amount for each present and future employee. By introducing future employees at their target hiring date, this also serves as a summary of your hiring plan and makes it easy to play with different scenaria.
It’s tempting to include non-payroll items such as training and benefits in the total cost to company here, but I find that it’s best to list them separately in opex — this way it’s easier to make adjustments on the budget for such items and to track expenditure against them in the future.
To calculate total cost to company it’s best if you use a separate “staffing plan” sheet where you list all roles organised by function including other information like cost breakdowns (gross salary, taxes, national insurance, etc) and ESOP allocations — you’ll need that separately anyway.
In an 18-month plan you may plan salary increases at some point, so don’t forget to adjust for that.
Sales and marketing
Young SaaS companies are fond of claiming that they do “no marketing at all” but it turns out that even organic traction isn’t exactly free. Chances are you’ll need a budget for marketing collateral, (e.g. paying for a video, printing brochures) going to trade shows, expensing trips or producing great content to attract potential customers. At seed stage you may be doing much of that on a shoestring, with blog content written by the founders and riding the startup conference circuit and social media for cheap publicity. Over time, these will become more expensive as you do them systematically and at scale, so don’t forget them and don’t underestimate them.
Depending on the mix you use, it may be useful to split this into separate items such as field marketing, content and business development. I try to split the items so that I have access to numbers that will help me calculate other things. For example, if I have measured that it takes us X dollars in content production to bring Y number of new customers, it’s useful to have this as a separate category so I can sanity-check the acquisition estimates by the amount we’re spending.
If you use any form of SEM/PPC, this should be a separate category within marketing, because you’ll want to use the amount spent as a parameter for estimating traffic/acquisition volumes.
A lot of things small and large fall under operating expenses. You don’t want to make a long list of trivial items, but you should at least split them up in a few categories so you can see where money goes and also track against these categories when you monitor your actual spend.
Fully loaded office costs including utility bills and services like cleaning. Approximate this as a flat fee or maybe make it grow by a fixed amount per month to simulate the fact that running an office becomes more expensive over time as you get more people.
Employee benefits and training budgets. This should include healthcare, lunches, travel expenses, training tools, etc. It’s a function of headcount so figure out a monthly or annual budget per head and make it a parameter.
Software and telecoms, including licenses for productivity and collaboration software (e.g. google apps, CRM, project management) that your team is using. The best way to model this is as a function of headcount. Don’t get too caught up into who uses what license — just figure out your average spend per person in tools and telecoms and assume it will grow linearly with people.
Legal and accounting fees. You probably already know what you pay for compliance purposes but it’s also important to provision for (a) extra legal costs for your next financing round and drafting ESOPs and employment-related contracts (b) bookkeeping costs rising as your monthly transactions grow to non-trivial volumes (c) auditing and tax advice. If you know when these costs will be incurred, add them on their respective months. If not, make an annual estimate and spread it on a monthly basis.
Recruiting costs. If you’re about to raise a Series A chances are you’ll be spending money on recruiters, sourcing services and job promotion. (hopefully you’ll be using Workable to save some of that money and effort!) Assume a % of new hires will require recruiters who will charge X months of salary for finders fee and add the recurring cost of subscription and job promotion services.
Travel expenses for corporate development. If you’re not in SF, assume you’ll be doing quite a bit of travel for investor relations and fundraising. Travel for board members, interviewing remote candidates and non-marketing trips also count here. These should not include travel expenses you have for BD/sales. If you’re in Europe, pay extra attention here. A low-burn seed stage company may need to cut one month of runway just to pay for the fundraising trips, so make provision for it. I use historical figures to calculate a dollars-per-day value for a person’s trip and then assume a few weeks of travel per year plus board meetings.
Equipment and capex, including office furniture, computers, etc. If you’re furnishing a new office, get an estimate from a similar startup that did it last year. For personal equipment, draw up an average cost per employee and add that cost with a formula whenever the headcount goes up in the time-series.
COGS and infrastructure
This is the cost to deliver and bill the product to the customer. In a SaaS company it will typically include three classes of items:
Cloud infrastructure costs which includes your servers, (e.g. AWS) storage, and third party services you use as part of the product such as mail delivery, monitoring, dns, etc.
Billing platform fees which includes transaction fees, foreign currency commission, refund charges, or whatever else they gouge you for.
Third party services that you are reselling/upselling from within your product. This includes costs that do not apply across the board (as per the first category) but instead can be attributed to specific transactions. In other words if the customer pays extra (in the form of an upgrade, upsell or in-app purchase transaction) for some service and you pay a provider part of that amount, list it separately here because we’ll want to keep track of GP margins for it. An example would be if you allow people to buy a custom domain through your software that, in turn, you purchase from another provider. Or if they upgrade to an HD streaming viewer that you pay higher streaming costs for.
This category is a common pitfall for early-stage startups because at small scale these costs are trivial, but as usage grows they can grown non-linearly. If you do a good job at developing the right product, average usage per customer will grow in all sorts of ways: storage used, queries, seats, features used, etc. You will probably be adding heavier and more demanding features, need more supporting tools, open up separate services, replicas, etc. Also, for low-price point SaaS, the billing transaction fees can cut into your profit considerably. Do not underestimate this category.
If you have early traction ($500k-$1m ARR) you may be able to calculate an estimate of cost per month per customer based on historical data. Assume it’s going to grow a bit and then make this cost a function of customers. If you don’t have enough data, make an estimate and then multiply this estimate by 3.
For SaaS companies with some kind of freemium or free version (not free trial — I’m talking about having regular customers that don’t pay) you probably need to do a more sophisticated analysis here. Figure out what free customers cost you and how many of them you have per paying customer. Then attach the cost of the free customers to the paying customers. This will also help you evaluate if the free version is helping or hurting your business.
You may have some costs that are expected and planned but do not properly belong in other categories and will not recur with any regularity. For example, we paid an architect and a developer to renovate and decorate an office that we will keep for years to come. Sometimes it’s best to keep such items in a separate line.
Incidentals and over-budget
Nobody sticks to their budget 100%, much less a startup where things can be wildly unpredictable. Add a budget item for off-budget expenses or excesses. Don’t make the mistake of making this a flat amount per month. Make this a % of your total spend. Experience will tell you how much you tend to deviate, but if you don’t have much experience put at least 5% there. Even big and stable companies deviate by that much on their budgets. The earlier stage you are, the more liberal you should be with this one.
There are two approaches to forecasting growth and revenues:
The top-down approach is to observe your month-on-month % growth trend and assume it will project into the future with diminishing returns. Then work out what this means in dollars. Then assume a growth curve for ARPA (based on historical trends) and work backwards to find the number of active customers it represents, adjust for churn and thus figure out how many new customers per month this forecast requires.
The bottom-up approach is to observe what drives new customer acquisition, in other words how many new customers you get per dollar spent in each channel, estimate how much you’d be spending per channel, add organic/viral multipliers to the result and work upwards to revenues from that. If you have precise and solid metrics for your sales and marketing process, this approach may be the best for you. This usually works great for companies that employ a lot of PPC or lead generation from inside sales forces where the economics can be clearly understood.
I personally prefer to use the top-down method as a primary model and then use any knowledge I have for bottom-up metrics as a way to sanity check and validate that the top-down model target is achievable. You can do the opposite, too. Whatever you do, try to both methods and see what they give you, use the results of one against the other to tune your estimates. You’ll be wrong so the objective here is to be less wrong than using only one method.
Churn and Upgrades
Don’t forget to factor in the effect of churn and upgrades into your model. Both are likely to grow if you’re doing a good job. Churn will grow because your product market fit is just now improving and many of your early customers were trigger-happy early adopters or signed up for something different than what you will have in 6 months. Upgrades will grow because improved features, improved adoption and usage will lead to upsells and moving to higher plans.
Even though your upgrades could be larger than churn — resulting in negative churn, a common thing for really good products early in their development — you should track and model these separately. Maybe you’re not losing money but you’re losing customers and you need to be able to calculate how many more new customers you need just to stay stable which affects onboarding capacity and other metrics.
If you don’t have good historical estimates for churn (you can’t have if you haven’t been around for 2–3 years at least) be very conservative with it, i.e. assume higher than your average today and higher than industry average. Churn plagues early stage companies with volatile product market fit, so assume the worst!
Cash flow and runway
Cash flow is not the same as MRR if you have a mix of monthly and annual contracts. Track MRR by spreading annual contracts on a monthly basis, but when it comes to cash, you need to calculate how much you tend to collect up front. In our case the % of annual contracts is relatively stable so I can predict how much cash we will collect on average as a function of the MRR additions.
To calculate runway, remove any non-recurring costs from the denominator, so that you’re estimating roughly on recurring basis. If you have significant one-off costs ahead of you it might be worth writing a formula that takes them our of the numerator in advance.
You’ll want to start fundraising with a few months of runway left, so look at where the company will be like when you have 6 months of runway and honestly think if you’ll be in shape to win the hearts (and money) of investors.
Your budget presumably assumes that you’re spending sizeable amounts on marketing to keep growing fast. This means that your “inflexible” or “maintenance” spend, the amount you’d need to sustain your customer base without growing too fast, is less than your total predicted spend. Knowing how much churn you need to replace and your CAC, you can calculate a reduced total spend that includes just enough marketing to maintain your MRR or grow very slowly.
How is this useful? It gives you the minimum stable burn rate (without axing the company) and thus the maximum runway that you have at any given point in time. If you’d be looking to optimise the timing of a new investment round or hold out during a period where financing is hard, this is maintenance runway is your grace period where you can extend your lifetime without huge sacrifice.
It also serves as a sanity check. A post Series A SaaS company typically uses its financing to grow aggressively, so there should be a point in time about a year from now where you can choose to tone down growth spending but can still survive. If you have an infinite maintenance runway sometime in your planning period, it probably means that you have planned correctly and your balance of spending between marketing and product is reasonable. If you don’t, this may mean that most of your spending is still going into product, in which case you need to decide if you are doing this deliberately.
Sanity checks and pitfalls
Now that we have a basic model with initial assumptions, the next step is to debug it. Even if you did a great job with your assumptions, when the variables combine and multiply over a period of 18 months, a slight overdose can take some numbers off the charts. Note that any VC worth his salt will be digging for such issues anyway, so better find them before they do.
To spot such problems we’ll calculate a few “sanity checks” at the bottom. Here’s a few that I use:
If you have a reasonable product-market fit, your average revenue per account should be expected to grow over time. One reason for this is that, presumably, you are improving the product during that time creating better fit with higher-value customers or adding pay-as-you-go features and third-party services. Another reason is that typically churn is higher on cheaper plans in SaaS, with bigger customers sticking around for longer. So, unless you have a situation where you are intentionally going for higher volumes on low-end pricing or you’re fighting a price war with your competitors, your revenue per customer should display some growth over time.
Fully-loaded, blended CAC
Add up all marketing, sales and customer success costs to calculate your CAC and check that it’s within a reasonable range compared to CLTV. (1/3 of CLTV is considered good) If you’re below $1m ARR, much of your early traction was coming from highly efficient organic channels, but as you scale you’ll have to use more costly channels and waste some money until you become efficient. This means you should expect to see CAC growing higher. If not, maybe your assumptions are wrong, or you’re not pushing for volume as much as you should. (spend more, open up new channels)
Channel capacity ceilings
Just because a channel yields you customers for a good CAC, doesn’t mean it will scale infinitely. Split up your acquisition estimate to different channels (if you can’t do it bottom-up at least do it top-down by assigning each a % of new customers according to your best estimate today) and see how many customers the model expects for each channel in absolute numbers. Does it make sense? For example, if you’re getting 5 customers a month today by going to conferences, can this be 50 a year from now? Are there enough similar conferences? What about 500?
Channel growth rate ceilings
Some channels may have very large ultimate capacity but increasing your spend in them cannot be done overnight. You can’t go from spending $5k to $500k in adwords in a month, even if you have the money. You can’t hire 10 new salespeople and get results from them in less than a quarter because they need to be trained and they’ll need 10 times as many leads as the one person you have today. Calculate the growth per channel (both % and absolute number) and verify month-by-month that there are no impossible jumps there.
Divide the number of new customer onboarded per month by the number of customer success/support you have on your headcount. Is it reasonable compared to what you do today? Customer service effort grows fast and new people will need to be trained and won’t be as effective when the team grows larger. Make sure you scale headcount and costs in line with revenue growth. What I do is set a number for onboardings/headcount and set our hiring plan such that we always get a new person in the team a couple of months before we hit that limit.
Same as above, but this time we’re looking at total active customers. Look at the % of customers requiring support at any given month so that you can calculate support issue volumes in the future and establish a max number of issues an agent can deal with per month. Check that the ratios stay in safe range throughout your 18-month plan.
Marketing budget per field salesperson
Is your marketing budget growing to accommodate your new sales hires? Divide your field marketing spend by the number of sales reps on the field. If it’s too little your sales reps will not be able to do enough. If it’s too much, they won’t be able to spend it and that means your revenue assumptions driven by that marketing figure will not be fulfilled.
I run some quick ratios like OpEx per head and revenue per head just to get a feel of how reasonable they are. These are numbers that you can compare to other companies (ask a few founders you know) to get a sense if you’re spending too much on OpEx and decide whether this is by choice or just bad planning. Another thing to consider on the per-head items is the number of people and infrastructure you have in place to serve them. Can you have a 20-person company without an office manager? Can you grow to 50 people without some kind of HR?
Per customer ratios
Similar to the above, run every figure about the business on a per customer basis and try to interpret the results. We pay the bank X dollars per customer? We have a person in marketing for every Y new customers we get a month? How does this compare to others? Does it look strange? Why is it so and what does it mean? What factors affect this? Maybe I should look into those cost items and re-evaluate them.
What’s life going to be like for this guy?
This is not strictly a ratio, but a way to look at roles and functions line by line in the context of a future scenario and think about how sustainable that future would be.
Here’s how it goes: I pick a moment in time lets say a year from now and go through every person in the company thinking things like “So, now we are generating 5,000 invoices per month including a few dozen problematic cases with frauds and refunds, we’re managing contracts with 7 suppliers, payroll for 40 people, audit closing dates are near in two countries and we’re about to negotiate a new term sheet. This guy is our financial controller and it would be his job to deal with all of this. Is it reasonable to assume he wouldn’t need to have hired an assistant already?” You get the idea. Is your one marketing designer enough to support 4 people in content marketing and $200k/month spend in online marketing campaigns that need creatives, landing pages, etc? Do you need a new product manager now that you have 10 more engineers?
The problem with forecasts is that it’s easy to just scale up the numbers on known models but forget that at 5x or 10x the size, the way we do the work today may not longer be viable. So we’re trying to imagine what the company looks like in the future, and what work would be like, in order to verify that we are not planning an impossible situation.
Even for a product with minimal seasonality, there are some times of the year when things slow down. December for example is almost like a half-month for new business in many SaaS products. There may be special seasonality around your product category on top of the overall seasonality of business customers. Also, if you’re selling B2B, weekends are likely to be off so any per-day calculations you do on ad spending, support staffing, etc will need to take account of that.
Normally, you can make a simplistic model that treats all months as equal and it could work out ok, because some other months are going to peak and the overall result will be similar. But it’s demoralising to miss you target on a month, even if you know that the target was inaccurate in the first place, so adjust if you can.
A less-obvious place to look for seasonality is hiring. You won’t be moving as fast with interviews in early August or Christmas, so if your plan assumes making offers to three people on January 1st, it will probably not play out exactly like that.
Similarly, don’t count on fundraising in August, so make sure your runway doesn’t clash with the holiday schedule of VCs.
Depending on your business and the timing of things, such effects may be trivial to the overall P&L, so feel free to ignore them. But be aware of them because in some cases it matters.
It’s only good if you can validate and improve it
Just because you modelled something, doesn’t mean it’s correct. The purpose of modelling is to make you think about the dynamics that affect every variable in your financial plan, so you can understand what drives it. These in turn allow you to explain your figures in the form of assumptions, which is something you can validate and tweak with research and historical trends.
In other words, you may not be 100% sure that your numbers are correct, but at least you’ll know why they are there and you can debug or fine tune your plan as you learn more about your assumptions.
I think the biggest mistake of inexperienced entrepreneurs is to think that “since we don’t really know, let’s just guess”. There’s a big difference between guesswork and an informed model. Both may be wrong, but the latter can be understood, fixed, and improved.
Some things in your financial plan will be very hard to estimate, and usually these are the things with the larger impact. For this reason, it’s worthless and a distraction to over-analyse small and trivial amounts. Sure, you can make a formula to adjust the cost of free beverages in the office for inflation, or spend a day with the IKEA catalog to figure out exactly how much you’ll spend on office furniture. But is it worth it when a 1% error on the growth curve can throw you off by a half a million?
My friend Christoph Janz has an excellent post about Parkinson’s law of triviality and how to avoid it in startup business plans which I recommend you read.
Target setting and VCs
VCs have this terrible habit of rejecting ambitious plans as unrealistic and then tell you that you’re not ambitious enough when you present them with a more reasonable estimate. Ignore them and make the plan that you think you want to run your company off of.
Don’t reverse engineer your plan to fit the numbers they want to hear — inevitably this approach will create a monster that doesn’t make sense. (and the same VCs who forced the creation of the monstrosity will be eager to point out the inconsistencies it contains, more proof that they are very clever and their MBAs were totally worth it)
Regarding targets, it’s always good to shoot at the edge of the achievable. I.e. a bit higher than you expect, just so slightly higher than the expectation that it will be still achievable and motivational. The financial plan helps you define achievable. You want to make targets for the sales team or for VCs? Jack it up a bit and the go through the sanity checks to verify that it’s not bonkers.
With so many tools to measure, experiment, and quickly iterate through the designs of cloud software products, it’s tempting to fly by instruments and “let the market show you the way”.
Engineers love structured and formulaic methods, so they’re naturally drawn to micro-improvement problems. The thing is, making something better only works if you start with something good. And good product design starts far outside the realm of the quantifiable. For this reason, it’s a bad idea to rely too much on iterative optimization during the early, formative years of a product-oriented startup.
Better is easy, good is hard
The biggest problem with micro-improvements is that they’re commercially introverted. They don’t drive real growth. Real growth comes from making a product that meets customer needs, telling a story that resonates, designing an engaging experience, building scalable distribution channels, making people happy, and getting them to talk about you. None of these things can be built by A/B testing or polishing funnels. Yes, they can be improved — but that can wait for now.
Click-through rates don’t write good stories
I’m sure The Economist employs great editors that write compelling headlines, but this rests on a solid layer of good reporting and analysis. If you tried to build a publication simply by optimizing for the click-through rate of the headlines, you’d probably end up with a tabloid.
There’s a common misconception that optimization is a series of micro-improvements to polish something into perfection. So, the more stuff you try, and the more things you optimize, the better your product gets. But it’s not about perfection through accumulation — it’s about choice. You optimize forsomething. You choose something that you sacrifice other things for, and this implicitly sets a design direction.
The dangerous mind-game of micro-optimization is that it consistently yields improvements in this or that metric that give a sense of progress. Setting your direction to chase whatever metric seems to be the easiest to push at the moment is an easy way to get trapped around a local maximum. You tell your investors that you’re focusing on the high-impact factors, but the truth is you might be going down a product design rathole.
Behaviour is a symptom, not a cause
It’s a very common mistake to optimize for proxies of an outcome, instead of for the outcome itself. We make recruiting software. Statistically, our loyal, paying customers share some common behaviours: they add our jobs widget onto their website, they post to job boards, send emails to candidates, and they invite their colleagues to schedule interviews and give feedback.
I can trick you into doing some of the above tasks with some big yellow buttons (we’ll A/B test the colour) that each say — actually, never mind what it says, we’ll A/B test that too. But does the correlation work backwards?The common behaviours are only symptoms of the collision of people who care about hiring with software that does the job well. A substantial and lasting increase in the number of people that engage with the product cannot be achieved by gimmicks.
Funnels are a useful model for troubleshooting flows and discovering hidden causalities, but it’s worth remembering that they are merely an abstraction of a segment of user experience, not the experience itself. You can’t compartmentalize user experience and drive it by min-maxing clicks and screens.
User behaviour is something to be measured, understood, and interpreted. Design follows the interpretation of behaviour but shouldn’t try to force it.
In a startup environment where resources, time, and attention are in short supply, it’s vitally important to pick one thing that you’re going to obsess over for at any one time and only optimize for that one thing.
Optimizing for acquisition is a good bet. It’s a wide enough goal to affect the design of everything that relates to product-market fit and marketing, but leaves some complex questions like pricing, churn, and cost-efficiency out of the picture for the moment. Focusing on increasing a % conversion on step 2 of some arbitrary funnel is too specific and too low-leverage to make a meaningful impact on your product development. Testing, measuring, and iterating are important tools for a product designer, but they’re only helpful if you ask the right questions.
The problem with insignificant battles is that they’re still battles nonetheless. It’s too easy to spend enormous amounts of time optimizing for something that won’t make much of a difference. To make it worse, it’s a sophisticated-looking task, so it feels like a good day’s work. We’re suckers for streams of small but frequent rewards.
Don’t get me wrong. It’s healthy to be analytical and make decisions supported by factual evidence. It’s smart to harvest quick wins by removing adoption blockers and refining critical flows. It’s beneficial to nudge users here and there. But you can’t build an entire product by throwing shit at a wall to see what sticks.
It’s been the most interesting time of my life. For what it’s worth, here’s what I think I learned in a year of building a venture-backed startup with 15 amazing people and a product used by companies all over the world.
Cheating product research
Making a good product is hard. You can’t fly by instruments, because at first, there aren’t any. You can get a basic sense of direction by building something you’d use yourself. This way you already have some idea of what’s missing and what would be awesome to have. Our major de-risking trick was to build something that we knew at least one person would buy: ourselves. I know it sounds like chickening out of product research, but hey, startups are hard so I’ll use any unfair advantage I have.
Far-sightedness versus near-sightedness
I was disappointed to find out that customers don’t buy visions and frankly they don’t give a damn for your revolutionary ideas about software. They buy things that work today, meaning features and immediate results. So you must take small steps along the path customers carve for you if you want to be relevant. But you can’t let them lead you completely, you still need to see far ahead and nudge everything towards a destination that exists only in your head. This is a hard, schizophrenic, mind-bending design process. One of the hardest things in my life today is the constant choice between being far-sighted for one thing and near-sighted for the other.
Growth versus optimisation
The geek inside me wants to optimise. It’s a comfortable space. Have something, tweak it to make it better. Be smug about it. When that something is 100 signups, best spend your time getting another 100, instead of improving conversion by 0.5%. In the beginning, growth is the only thing that matters.
Measure everything. You never know when it’ll come handy. But it’s easy to fall into the nerdy trap of trying to interpret everything to soon, and improve every single metric. This isn’t helpful. Find out the one thing, the one metric, that impacts your growth the most and throw all your weight on it. Every three months, re-evaluate what is the one thing that matters most. With all your team. Features, marketing, time, every single person in the team pushing the biggest lever for growth. You’re not strong enough yet to be pushing two levers at once.
I learned to not shy from asking people to pay for our stuff. If you think they won’t, then you’re building the wrong thing. Build something people want to buy. When you find a few of them, make them very happy. It should be bloody obvious, but in the startup twilight zone it isn’t, and in your early days of despair you’ll take anything for “traction” even if it’s a random bunch of freeriders sending you in the wrong direction. Ignore them and listen to the folks with credit cards in their hands.
Selling versus explaining
Some people think I’m a good salesman. I’m not. I just cheat my way out of having to sell. Here’s how to cheat in sales: (1) Make something awesome that makes a ton of sense, and (2) Explain how it works and why it makes sense. Tip: the first part is the tricky one. All other methods of selling are too fragile to build an entire company upon them.
A self-evident future
The best way I’ve found to explain what we’re doing to investors is to walk them through a self-evident vision of the future. Help them think about why the current status of our market cannot remain the same in 5 years, and what it will probably look like when it changes. Then I only need to show them that the future looks remarkably like what we’re building today. How do you make your vision self-evident? If you ask this question, then your problem is not the pitch, it’s that you don’t know why you’re doing what you’re doing.
Inane argument killer technique
VCs will find a million ways to explain to you why this will never work, or it won’t work so well, or it will work but it won’t be scalable, while insisting that it must have “a mobile strategy”. I have discovered a technique for insta-killing these inane discussions. It’s called “the graph with actual customers and revenues that grow faster every month”. In my experience, until you have a (real) graph like that, don’t even bother. When you do, you’ll never grow tired of hearing “please go back to that slide”.
Startup misery folklore
80-hour weeks, all nighters, eating ramen noodles, maxed out credit cards, sleepless agonising nights tortured by 3-week runways.. I say that’s bollocks. Build a happy workplace, strive to be efficient, have a sensible business model and pay your people enough to live comfortably. What makes you a startup is that you’re innovating and building something from scratch. Having a shitty lifestyle is just a shitty lifestyle no matter how much hipster-sauce you dip it in.
No compromise on people
You want to disrupt an industry by going up against companies whose stationery budget for the year is bigger than your seed funding. You’re not going to do it with average people, so this is one area you can’t compromise. Making the company attractive to bright folks, actively searching for them and inspiring them to join and stay is probably the single most important job of a startup CEO. If you get that right, everything else becomes so much easier.
Doing things differently
Don’t. There are so many things to do, and if it’s your first time running a company, it’s tempting to do it “my way”. Formulaic aspects of running a startup are formulaic because they work well this way. By all means, you need to do some things differently, but differentiation has an impact if it’s intentional, calculated and in small, precise doses. Doing everything differently looks more like you’re doing random shit and you have to ask to ask yourself if that’s indeed what you’re doing.
Drawing first blood
Investors, advisors, employees, your network.. if you’ve done a good job they are all people that can and will support you, put more power behind your every thrust. However, you have to draw first blood. Take the first risk, open the first door, sell the first client, articulate the vision, ask for what you want, be on the frontline. You’re the poor guy with the machete at the head of the expedition clearing up a path through the jungle. Don’t expect anyone to do this for you or on your behalf.
Strength and determination
Running a startup is a constant struggle to persuade all sorts of people to ally with you and trust you. Investors, customers, employees. They don’t care if you’re having a “dark hour” or you’re being crushed under the burden. Nobody cares to ally with the weak and the wounded. No matter what, you must appear strong, determined, going places. This can be really hard sometimes, so it helps if you’re a little crazy-determined to start with.
I learned how to say no. It’s a survival skill.
Yes, it’s hard, but so are all amazing and rewarding things and it’s only possible if you enjoy it. And I am enjoying it, dammit. Every moment of it. Even the really crappy ones.
Don’t look for evangelists among people who think your product isn’t worth $19
We used to offer a free plan for light use of Workable. A few weeks ago we eliminated it. In its place we created a very affordable plan at $19/month.
There has been plenty of interesting debate on the merits of freemium for consumer products. I think our experience offers an additional perspective, from the standpoint of a b2b SaaS company.
So I’m sharing our reasoning here, for what it’s worth.
Limiting design choices to things that can scale infinitely at near-zero cost is a recipe for making mediocre products.
Free users are not really free, even for a digitally distributed product like Workable. They typically outnumber paying customers by a factor of 10 or more. As Workable started becoming more and more popular, we realized that this isn’t going to scale. We know that a few months from now we would start getting crushed under the weight of our own success.
The best features become impossible to incorporate in the product. You can’t add any high cost-per-use features that require third-party technology or heavier operations. For example, we wanted to offer our users a better viewer for resumes, advanced parsing capabilities, better search algorithms, and so on. The cost to offer it to thousands of users for free became prohibitive.
We are simply not willing to limit ourselves, our product and our customers to mediocrity, so we found a low-end pricing that’s affordable for light users but permits us to build the product they deserve.
Some would say why don’t you offer the better features only to paying customers? But this is not as simple as it sounds. Do you create a gimped product that’s lacking half the functionality? Do you maintain a good version of a feature and a poor one at the same time, with all the complexity this adds to the product?
What about support? Do you ignore non-paying users, offer them no support or bad support? This will only lead to frustration and bad customer experience for a huge volume of people, eventually destroying the product’s reputation.
The idea that free users generate free marketing is self-defeating.
It’s easy to think that the more users you have, the more free marketing you get.
More accurately, it’s happy users that create free marketing for a product, and we haven’t found a magic way to make free users happy at zero cost.
Don’t look for evangelists among the people who think your product isn’t worth $19.
Users who can be made happy enough to create good word of mouth about your product, are probably the same people who would pay for it already.
Free users are notoriously hard to please. They want something for nothing. They probably don’t like your product too much and they’re not getting a lot of value from it. (if they did, they wouldn’t mind paying $19 or $49 for one of your inexpensive plans)
Free users will get the worst possible experience. They see the most minimal version of your software, you have put limits on their use of good features that cost you money, and they’re last in priority for customer support.
Maintaining a second-class version of our product, just for the sake of having a free plan, would only create a mass of under-served customers, ultimately hurting our brand.
On the internet, your $1,000/month user doesn’t have a stronger voice than the guy who never paid for software in his life.
When someone tweets “this product sucks” readers can’t tell if this is coming from someone who got the good version of the product or the bad, unsupported one. Oh, and for every paid user you probably have 10 or 100 free ones. They will be louder, by sheer numbers alone.
Is it really good marketing to make a mediocre impression to a lot of people? We decided we would rather have fewer people experience our product so that we can give them a product worth experiencing with the quality, features and support that we want to define our brand.
Customers who expect to get real value out of your product are the only customers you have a chance of making happy.
There’s a world of difference between “free” and “very affordable”. It forces an honest decision. Will I really get value out of this product, enough to justify paying for it? If someone is not getting $19 worth of value out of Workable, then we’re not solving a real problem they have.
We only want customers that we can hope to make happy and we don’t see a sustainable way to do this for non-paying customers.
Let’s be honest. Whatever you do, supporting an overwhelming host of free users does have some cost, even if you manage to keep it low. Guess who will pay this cost? The paying customers, of course. We don’t think that’s fair.
We know that not everyone will like and need our product enough to justify paying for it, and that’s ok. We just want to spend all of our effort and resources serving those who do.