How does proper management of your financial data improve your business outcomes?
You can make data-driven decisions – startups often want to do an ROI analysis on capital spend or marketing campaigns—and you need concrete financial figures for that. You need robust financial data to understand key metrics like CAC, LTV, etc., or you’ll be justifying spend based on hunches.
You’ll have better visibility – your P&L is not the full accounting picture, you need to look at cashflow, performance to plan, and all of the numbers that roll up to P&L. For example, if a big software spend is coming up, you need visibility into that for cash flow purposes, and to forecast for it.
Your finance and accounting numbers become more meaningful – don’t just think of finance and accounting as the monthly close package. The real magic happens in the guts of the numbers. Understand churn, your pricing, how you compare to the market, and your product’s value to customers.
Why is good financial data important to investors?
Show investors you have a handle on your business – if you report on CAC or churn rate, show investors you know what the numbers mean to your business to inspire confidence. Having some data collection at an early stage is important to show potential investors you understand your business at the level that’s most important to them: the bottom line.
Improve on the bottom line – for all of the reasons that financial data is important to your business, it can improve your performance and lead you to a more favorable financial profile for investors.
At what scale does it become especially important to capture and manage financial data?
Get data-capture processes in place as early as possible – you can “wing it” as you onboard your first few customers and your back-office processes are still being worked out. Trying to get mature insights at that stage is like trying to build a racecar as it’s zooming around the track, but still try to collect data so you don’t have potential value going out the door.
Professionalize in Series A – once you’ve proven your model works, it’s time to start taking a long, hard look at your processes and what data you have. You’ve likely done some financing already and have a rough chart of accounts, but your financials couldn’t be presented to a bank. When you raise your Series A or beyond, investors expect you to know your business and be able to present your financials professionally and in accordance with GAAP.
Show operational excellence by Series B – by now, investors expect you to have a firm handle on your finances, leverage them for growth, and show operational excellence. Your financials should be locked in tight at this point with a thoroughly established chart of accounts and proper coding of all your income and expense accounts, an accurate balance sheet, and a clear cap table.
If you’re bootstrapping, get the basics in place before “hockey stick” growth – it doesn’t necessarily require a massive investment in tools, but you need to be sure that the ones you’re using are scalable and your processes are ready to go.
What type of reporting should finance generate regularly?
Finance should own all KPIs – finance is already trusted to report financial data, and they use the KPI data in their models. Finance should report on and answer questions about customers, vendors, and operational performance, and tie that back to the financial statements. This is important when establishing the source of truth for all your metrics (read more on this topic below).
Finance should explain through reporting – if, for example, there’s a discrepancy in a COGS account, then the next step should be to create a report that explains the variance. Finance should provide concrete value to the rest of the company with analysis and reporting. Don’t try to cover up or hide variances or discrepancies. Instead, be prepared to understand and explain them.
What are the different financial tech tools for capturing and managing data?
| Core GL/ERP | ||||
| What it is | A GL is your basic accounting and finance platform – e.g. QuickBooks, you can use it to log all your transactions, upload budgets, report, and run the key functions of accounting and finance. These software packages are typically limited to pure financial analysis. An ERP is more robust – e.g. Netsuite, Enterprise Resource Planning tools are a type of software that organizations use to manage day-to-day business activities such as accounting, procurement, project management, risk management, and compliance, and supply chain operations. | |||
| When you need it | You need basic accounting software on day one, as soon as you open a bank account. You don’t want to be operating out of a shoebox full of receipts when it comes time to file your first tax return! | |||
| When should you move up to an ERP | When you can’t meet business and reporting requirements with a GL, that might happen if: – You have complex data or accounting needs, e.g., international entities, many different entities, tracking your sales pipeline early, or demanding reporting – You want to signal operational excellence to investors But make sure you have the resources to handle an ERP. Consider the following: – All in, you’re going to spend $200K+ just to set up an ERP – You need full buy-in from your full management team – 60% of ERP implementations fail. | |||
| Integrations | Everything. To whatever extent possible, your finance and accounting system should have direct connections with your CRM, bank tools, bill payment tools, etc. | |||
| Accounts Payable | ||||
| What it is | Tools to track AP – e.g., Bill.com, you can set up approval flows, automate payments, and track invoices. | |||
| When you need it | As soon as you have QuickBooks. AP automation tools are inexpensive when compared to the efficiencies of tracking and having a single location for AP management. | |||
| Integrations | Integrate with your accounting system and your purchase order system if it’s separate. | |||
| Accounts Receivable & Payment Processing | ||||
| What it is | Tools to help you collect on receivables – e.g., AR Collect, Tresorio, HighRadius. Some orgs have whole teams dedicated to sending statements, making phone calls, and tracking promises to pay. With just a bit of automation, you open up bandwidth to focus on the most egregious offenders – the oldest receivables and the biggest invoices. 80% of what you do will be automated so you can focus on the 20% of exceptions that are most critical. | |||
| When you need it | As soon as you start sending invoices. Cash is king, and when clients are slow to pay, you’re basically giving them an interest-free loan. | |||
| Integrations | It should seamlessly integrate with your accounting software. There, you can mark invoices as paid and get real-time visibility into your AR, for more accurate cash flow forecasting. | |||
| Tip for leveraging | Give your clients as many payment options as possible. Make it easy for them to pay you by ACH, Credit Card, or check. | |||
| Quoting | ||||
| What it is | Tools that make it easy to create quotes, statements of work, and MSAs – use them to set up quoting parameters for sales to work in, so that you only deal with exceptional quotes manually. Typically, you use your CRM (e.g., HubSpot or Salesforce). Use a purpose-built CPQ tool if you have a big sales team or variable contracts. | |||
| When you need it | As soon as you have a sales team. Use a purpose-built CPQ tool if you have a big sales team or variable contracts. | |||
| Integrations | Integrate with your accounting software and project management tools. | |||
| Tips for leveraging | – Don’t make your salespeople become technical or administrative experts to quote a customer. Spend the time to set pricing rules, max/min, and eliminate possible mistakes. – Cover as many exceptions as possible, because sales will find a million of them when they’re trying to win a client – Try to make all your quotes and SOWs digital with electronic signatures and automation. Have everything in your systems, including rate cards, margins, etc. | |||
| Contracts/Documentation | ||||
| What it is | One place to store and manage customer contracts – e.g., SharePoint and Google Docs. It’s a tricky problem because people aren’t disciplined with file locations, create duplicates, and have poor change management. Software can help grant visibility for as many people as possible, with change controls. | |||
| When you need it | Try from day one—knowing you’ll make mistakes. | |||
| Integrations | The best setups build contracts into sales workflow through your CRM or ERP. | |||
| Tips for leveraging | At larger organizations, the CIO or CDO should own contract management; it’s part of your information management strategy. | |||
| Business Intelligence | ||||
| What it is | Tools for analyzing large amounts of data – e.g., R, Python, and SQL. This is the data gold mine of business! It can provide you with: – Descriptive analytics – starting to glean information from data (i.e. being able to track and explain data in your current environment) – Diagnostic analytics – attempt to answer why historical events happened – Predictive analytics – modeling future activities – Prescriptive analytics – use data identify and impact future trends | |||
| When you need it | As soon as you can pull off BI and hire an FTE who’s equipped to handle it. | |||
| Integrations | Should be the lifeblood of your organization, connected with every tool that creates data. | |||
| Tips for leveraging | – Your tool will only be as valuable as the skill of the operators using it – Drag-and-drop tools are available but very expensive and still require a knowledgeable operator – There’s a lot you can do in Excel before you’re ready for a BI tool | |||
How should you think about building connections between tools in your financial technology stack?
Start by evaluating your setup with an audit of your existing people, processes, and systems – before you connect anything, understand how data and information moves through your entire organization. Watch out for roadblocks where someone is tracking data in a single Google Doc or spreadsheet on their laptop. You need to know what you’re mapping and moving over before you start making connections.
Most systems can be connected, but ask yourself if they should be – even if the software isn’t designed for APIs, there’s typically a way to get the data to flow out of it. You can’t have isolated, siloed data and expect to be able to report on it effectively—but don’t just connect everything, there are pitfalls and concerns to consider about every connection.
How can you use unique customer IDs to link data in different systems?
Create an identifier at the first touchpoint (usually your CRM) – let your first touchpoint create the UUID and follow the customer all the way through. Other systems might have different identifier numbers, but they should all be related back to the first touchpoint ID. The thought of data being re-entered in multiple systems should keep a data-minded CFO up at night—the ID needs to follow customers from lead to prospect to onboarding, invoicing, and eventual churn. This is foundational to understanding CAC, LTV, etc.
If you don’t have a unique ID flow, you’re in for a lot of manual work – without a global identifier for each client, you’ll end up resorting to alternate methods to tie information together between systems. This could involve cumbersome matching tables to track and assign the data by mapping IDs to one another or trying to match records on other error-prone data points like customer name or phone number. This can be intensive work, and generally requires a lot of ongoing data cleanup.
How can you ensure data completeness and quality?
Have a single source of truth – understand the sources of your data and consolidate it (either physically or with rules on how it’s used). Create an authoritative system of record for customer data—usually, this is your accounting system, but it could also be your CRM.
If different systems have different data, figure out why and rectify the variance – you may need to remove duplicates, clean your data, or standardize it to make sure it’s ready for easy analysis.
Getting to a data warehouse is ideal for hygiene – a data warehouse will help you aggregate data from different sources to get cleaner reporting to investors.
What should be your single source of truth for data? How should you use it?
| Approach | Pros | Cons | |||||
| GL is source of truth (pipe in data from CRM) | – Very clear source of truth with a low chance for duplicates | – QuickBooks is limited, so you need to have data definitions and data dictionaries – You may need different information that isn’t stored in your accounting software (e.g. Quickbooks may have a billing contact, but not an operations contact) – You need to maintain an understanding of which tool has the most current version of each number – You need to overwrite false or outdated data in 2 places – If your project management and accounting software aren’t linked, you have to overwrite data in two places – If the salesperson doesn’t input info in the CRM, you’ll miss it in your GL (e.g. the sales tool may have the point of contact for who signed the contract, but this is likely a different person than the billing contact) | |||||
| CRM is source of truth (pipe in data from GL) | – Very clear source of truth with a low chance for duplicates | – You need to maintain an understanding of which tool has the most current version of each number – You need to overwrite false or outdated data in 2 places | |||||
| Bi-directional link between GL and CRM | – You never have to reenter information twice – Updates in a single place will be transferred over | – You can end up overwriting good data with bad data – Higher chance of duplicate information – The links and rules can be difficult to set up | |||||
| ERP | – All of your data is in a single repository | – This is the hardest setup to get to. It requires advanced rules to prevent undesired overwrites. | |||||
Selecting a single source of truth is a joint decision – it’s not just finance, everyone needs to be accountable and have rails along the way. If the systems aren’t talking to each other, you have to rely on whoever’s entering the data and running the report to follow the rules.
Some things to look out for in your single source of truth:
- Poor data quality – garbage in, garbage out. Are the people entering data inconsistently? The quality of data goes down if you’re typing into text boxes with misspellings, fat fingering, different phone number formats, etc. Also, data formatting can be inconsistent if you’re piping it from different systems.
- Missing metadata – if you’re not piping in full records, you may have missing metadata.
- Difficult to access data – you’ll have to pull data from sources that may not be designed for exporting data. Access is a common problem with proprietary tools that try to keep you in their software.
When do you need a data warehouse, data lake, or data mart?
Start with a data lake – it’s a place to throw all the data you have, as is. Data lakes are a solution when you know you want to save and use data, but might not have a plan for how yet. The reason you don’t stick with a data lake forever is that running queries on these huge databases can be really slow, which makes getting real-time information difficult.
A data warehouse requires more organization – in a data warehouse, you’re taking data extracted from your transactional systems, and giving it some structure or modeling it before storing it. It allows for cleaner data with quicker and smarter querying.
A self-serve data mart isn’t feasible until a company is extremely mature – it’s difficult to make all of your data readily available to internal users in a data mart, but if you can do it, an organization with total democratization of data is powerful.
How can good data hygiene improve your company’s valuation?
Data fluency inspires confidence – you need to make it clear that you understand your business. If you have to wrangle data for a week every time there’s a question from a potential investor or your board, you’re not inspiring confidence.
It’s correlated with increased automation and cost savings – if you automate, decrease the number of manual errors, and decrease the amount of time cleaning data, you can realize significant cost savings.
What are the most important pieces to get right?
Don’t take your data at face value – understand what it means, where it’s from, and where it might fall short—don’t even talk about the data until you understand all of the systems.
Understand data entry points and your source of truth – understand the end-to-end process for customer data, and know what the benefits and negatives of each data source along the way are.
What are common pitfalls?
Being complacent with the first piece of data you see – don’t just trust your data implicitly. You could have duplicates or garbage data in there. The most common pitfall is not knowing what you’re doing, getting access to data, and reporting on it immediately.
Skipping ahead to complex reporting before you have clean data – getting to data maturity is a process. You need to master gathering the data, understanding its meaning and validity, cleaning it, and then it can become descriptive and informative. Don’t try to jump in with complex tableau dashboards that may look nice, but don’t give an accurate picture of what’s going on in the business.
Responses