Exasol — Analyse this

Exasol was founded in 2000 in Nuremberg, Germany, and has developed a high-performance, in-memory database designed specifically for analytics. This software enables companies to access and analyse data at high speed and scale. Exasol serves a broad range of market sectors and has customers ranging from small and mid-market businesses to large enterprises. Applications for the software include financial reporting, data visualisation and predictive analytics. The company has 156 employees in Germany, the UK and the US. Exasol listed on the Scale segment of the Deutsche Börse on 25 May 2020, raising €48m (net) from the issue of 5.1m new shares and the sale of 285k treasury shares at €9.5 per share and placing a further 3.8m existing shares.

Prior to the IPO, Exasol was self-financed through a combination of shareholder loans and bank debt. The company is keen to accelerate investment in sales and marketing in order to drive the growth of the business and has earmarked c 70% of the IPO proceeds to do this. The intention is to focus initially on developing the business in the UK and the US, where it has already had some success. In the longer term, further geographic expansion is desirable. The company is also considering M&A, and could use c 10% of proceeds for bolt-on acquisitions that support its market positioning and broaden and accelerate adoption of its software in specific verticals or analytical use cases.

Exasol’s analytics database software provides a data warehouse for structured data. The diagram below shows how Exasol’s software interacts with other applications and sources of data.

The main features of Exasol’s technology are:

The combination of these features results in the fastest analytics processing in the industry. This has been confirmed in tests against the industry benchmark TPC-H. TPC-H is a decision support benchmark that consists of a suite of business-oriented ad hoc queries and concurrent data modifications. The performance metric reported is the TPC-H Composite Query per Hour (QphH) for a given database size. The price/performance metric is expressed as $/QphH for a given database size. All TPC benchmarks are transaction processing tests undertaken by the Transaction Processing Performance Council (TPC), an industry body mainly made up of computer system vendors but also including some database software vendors. The chart below shows Exasol’s performance against two unnamed peers. Exasol clearly leads in terms of the volume of queries handled and the cost per query handled.

The company also reports that it earned best in class scores for customer service according to the Dresner Analytical Data Infrastructure Market Study 2020,1 and was a leader for both Customer Experience and Vendor Credibility for 2020.

The product is offered as one of three versions as per Exhibit 4. The Community Edition is often the starting point for organisations to trial the software for a minor data analysis problem before expanding usage to more critical use cases, when a paid licence would become necessary. Exasol One tends to serve SMEs or individual departments within organisations, and can be the seeding ground for larger businesses that may go on to upgrade to an Enterprise Cluster.

Exasol’s R&D department is based in Nuremburg and at the end of April totalled 38 employees. Five teams are working on the core database engine, each with a broad specialisation in several of the modules within the database and the ability to develop features autonomously. A separate integration development team integrates BI tools and big data systems and enables advanced analytics and data science in the database. Nearly all integration projects built around the proprietary database kernel are developed as open-source code to benefit from the contributions of the growing Exasol user community.

Exasol makes the majority of sales via its direct sales team (81% in FY19). It also has a partner network consisting of system integrators, business intelligence and analytics consultants, technology partners (eg Alteryx, Tableau, Yellowfin) and value-added resellers (VARs), which contributed 19% of FY19 revenues.

In recent years, the company has been relatively hamstrung in terms of increasing investment in sales and marketing. It is planning to use c 70% of the IPO proceeds to invest in additional sales and marketing headcount and to raise brand awareness. As part of this, the company wants to invest in channel managers to develop this route to market.

The company has sales teams in the DACH region, UK and US. At the end of April, the company had 52 employees working in sales and marketing and is planning to increase this substantially, particularly in the UK and the US.

The company estimates that it had a customer lifetime value (LTV) to customer acquisition cost (CAC) ratio of c 21x in FY19. This is calculated by dividing the estimated lifetime value of customers signed up in FY19 by the sales and marketing cost incurred over the same period (ie €3.7m LTV per customer/€173k cost to acquire each customer). As a rule of thumb, for a SaaS company an LTV/CAC of 3x is deemed to be the ideal ratio. Exasol’s high LTV/CAC ratio implies that increased investment in sales and marketing would have a significant impact on revenue growth.

At the end of 2019, the company had 175 customers, with 36 new customers signed up in 2019. Customers are active across many end markets, including technology, finance and insurance and retail (see Exhibit 5 for examples).

During FY19, 85% of revenues came from the existing customer base and 15% from customers signed up during the period. The company is focused on upselling to existing customers via three routes (in FY19, 21% of revenues were from upselling):

Data growth has exploded in recent years: IDC forecasts that data will grow from 33ZB in 2018 to 175ZB in 2025, a CAGR of 27%. The challenge for companies is to make use of the data they produce to gain useful insights into their business and to aid in decision making. With the ongoing trend for digital transformation, companies are increasingly becoming tech-enabled. Many companies now rely on technology to provide their services, for example online retailers, travel companies, comparison websites, sharing economy businesses such as Uber or Airbnb, and fintechs. These businesses produce vast quantities of data from diverse sources: customer transaction data, data from ERP systems such as inventory levels, social media, machine-generated data (IoT), emails, etc. In order to run their businesses efficiently and make the right decisions, the companies need to be able to manage the data they produce and analyse it in a number of different ways.

A database is a crucial partner to a software application. Each application, such as an ERP or CRM system, generates data that needs to be stored, updated, retrieved and deleted and it does this in a database. The software that controls the data in a database is called a database management system (DBMS) and it provides the ability to search, select and store data. This data can be structured (such as transaction data for a purchase) or unstructured (such as the contents of an email or a social media post). This results in databases being broadly grouped into two categories:

Exasol has developed an analytics RDBMS. An analytics database is a database management system that is optimised for business analytics applications and services in contrast to operational database systems, which are designed to create and update records in real time. Customers can store hundreds of terabytes of data in the Exasol database (a type of data warehouse) and then provide data access and processing capabilities to multiple analytical applications across the business.

The database software market has a small number of incumbent players generating the majority of revenues, much of which is from maintenance of their large installed base. In 2018, the five largest players (Oracle, Microsoft, IBM, SAP and Teradata) made up close to 90% of DBMS revenues (source: Statista).

According to Gartner, the DBMS market was worth $46bn in 2018 (+18% y-o-y), with cloud-based revenues of $10bn making up about two-thirds of the growth.

There are numerous newer companies that are focused on non-relational database software and cloud-first deployment, as well as the large public cloud service providers (CSPs: Google, Amazon, Microsoft and Alibaba). The large CSPs have developed DBMS software in-house, making abundant use of open source software as well as acquiring interesting start-ups, and unsurprisingly they dominate the cloud-based market. Amazon publicised the fact that it transitioned all of its own applications from Oracle databases to its own. Microsoft has an advantage in that it was already a large DBMS player before it developed its Azure public cloud.

Exasol tends to compete most often with the cloud data warehouse players, with the main competitors being Amazon Redshift, Google BigQuery, and Snowflake. Snowflake is a US-based private company that offers a cloud-only solution. The company is very well funded ($1.4bn raised to date, with the most recent round in February raising $479m at a valuation of $12.4bn). Exasol’s key competitive advantage is the speed of its software, which results in a very low total cost of ownership (TCO) compared to competitors.

Management board

The management board consists of three directors:

The supervisory board consists of four directors:

The company raised gross proceeds of €51.2m from the issue of 5.1m new shares and the sale of 285,000 treasury shares. Net proceeds totalled €48.1m. Of the pre-IPO issued shares, 24% were placed in the IPO (285k treasury shares plus 3.82m other shares from existing shareholders). The table below shows the shareholder structure pre-IPO and the expected shareholder structure post-IPO. Shareholders of 10.0m shares, including the Supervisory and Management Boards, are subject to a 12-month lock-up and shareholders of 2.2m shares are subject to a six-month lock-up. The free float is c 45%.

The company has outlined its planned use of proceeds, with the vast majority targeted at sales and marketing (70%/€33.7m), and 10%/€4.8m each for R&D, M&A and to pay for shareholder loans and the staff share scheme (this is the first priority).

Exasol offers its software via three different delivery models: on premise, hosted via the ExaCloud, or public cloud (including AWS, Azure and Google Cloud); and two licensing models: perpetual licence with maintenance contract or subscription licence. The company has increasingly shifted to a cloud-based, subscription model in recent years, but will still sell perpetual licences to companies that require one. The ExaCloud provides a managed service for customers who do not yet feel comfortable moving to the public cloud, or are not able to for regulatory reasons. Subscription licences are sold as a one- to three-year licence, with payment upfront. The company also generates revenues from services such as installation, updates, upgrades, administration, monitoring, incident management and training.

In March, the company undertook a global restructuring to prepare for the economic impact of COVID-19 disruption. It reduced headcount by 15% and introduced kurzarbeit (short-time working) from 8 April in its German subsidiaries, reducing working hours by 20% for the majority of the German workforce. The plan is to maintain those working hours for three months before reviewing. The company plans to partially compensate for the loss of remuneration by granting stock appreciation rights (SARs).

We note that Exasol reports under German GAAP rather than IFRS. See Exhibit 7 for the Income Statement.

We show in Exhibit 8 the revenue progression of the business since FY15. Adjusted revenues exclude a perpetual licence signed at the end of FY17, instead treating it as a subscription licence in FY18 and FY19, to show the underlying level of growth. We show the split of revenues between licences, services and other in Exhibit 9. The company generates a high level of recurring revenue, with annualised recurring revenue (ARR) at €17.6m at the end of 2019. The company started adding sales and marketing headcount from the end of FY18, which resulted in 50% growth in ARR in FY19. Recurring revenue of 70% in FY19 was made up of the combination of cloud-based subscription licences, on-premise subcription licences and maintenance contracts (with a roughly equal contribution from each). Customer churn has been relatively stable over the last four years, ranging from 4.0% to 5.5%, and was 4.0% in FY19.

The majority of revenues are currently generated in the DACH region, but with investment in headcount in the US and UK, we expect the contribution from DACH to reduce in relative terms (see Exhibit 10). In FY19, revenues from the US increased 94% y-o-y.

The company notes that although it is physically located in only three countries, it has customers in approximately 30 countries. There is some revenue concentration, with the top customer generating 11% of FY19 revenues and the top 10 contributing 35% of revenues.

Cost of materials comprises fees paid to partners, costs of hardware and hosting fees (for the ExaCloud). The company calculates opertating gross margin (revenues plus other operating income less cost of materials), which it views as an important metric as it incorporates all variable costs and represents the contribution availabile to cover the fixed cost base. The company has an operating gross margin of c 90% (see Exhibit 7). We note that the cost of staff providing services is not included within this calculation; the company estimates that this would reduce the operating gross margin by c 5pp.

From FY15 to FY19, the company saw variability in the EBITDA margin. Adjusting for the perpetual licence in FY17, the adjusted EBITDA margin was 14.5% in FY17 and 15.8% in FY18. In FY19, adjusted EBITDA excludes €9.77m of one-off costs, mostly the accrual for employee incentives (see below). Even after this adjustment, the margin was negative. This was mainly due to the 54% increase in underyling staff costs year-on-year as the company hired in the UK and the US (year-end headcount increased from 108 at the end of FY18 to 149 at the end of FY19).

The company has a stock appreciation rights (SARs) scheme in place and at the date of IPO, had issued 569,000 SARs. These give the holders the right to participate in the growth in value of the share. The IPO triggered the vesting of the SARs; for each SAR an employee is entitled to receive a gross amount equal to the difference between the share price and €1. The remuneration is due to be paid in three tranches: 30% 12 months after the date of the IPO, 40% after 24 months and 30% after 36 months. In FY19, the company accrued €7.9m for these within staff costs, included within ‘other provisions’ at year-end. The management team were also issued SARs calculated on a different basis with slightly different payment terms (the first instalment of €2.4m is due four weeks after the initial listing, the second instalment is due 27 months after listing). Prior to the IPO, certain shareholders transferred 881,794 shares to the company to be used solely to fund management SARs in future years. 285,000 of these shares were placed in the IPO and the €2.7m gross proceeds will be used to fund the payment of the first instalment of the management board SARs.

The company is also planning to implement either a stock option plan or a phantom stock programme.

The company is not currently in a position to pay a dividend as it does not have high enough distributable reserves. We expect that in the short to medium term, all profits will be reinvested in the business.

The company capitalises development costs: €1.4m in FY18 (7.9% of sales) and €1.8m in FY19 (8.4% of sales). Working capital is helped by upfront licence payments; deferred income stood at €4.2m at the end of 2019.

To date the company has been self-funded, with a mixture of bank overdrafts and more substantially with loans from shareholders. At the end of 2019, the company had a net debt position of €1.58m plus shareholder loans totalling €3.14m. In April 2020, a further €1.2m was lent by shareholders. All shareholder loans are due to be repaid after the IPO (totalling €5.6m including accrued interest at the end of April 2020).

As consensus forecasts are not yet available, we look at valuation metrics based on financial performance in the last reported financial year (LY) and compare Exasol to two groups of peer companies: US SaaS software companies and German software companies (which are a mix of SaaS and traditional software models).

We note that the US SaaS companies are much larger in size, both in terms of market cap and revenues. They tend to report recurring revenues at a higher level than Exasol, anywhere from 80% to 100%. As Exasol shifts more of its business to subscription licensing, and sells more via the channel (so it does not undertake implementation work), it should see its recurring revenues increase as a percentage of revenues from the current c 70% level. Of the 39 US SaaS companies included in the averages, 12 were loss-making in the last reported year.

The German software peers are all profitable at the EBIT level. Those with a market cap sub-€1bn grew revenues at an average of 16.6% last year, compared to Exasol’s 22% growth, and are forecast to grow on average 13.8% in the current year. On an EV/sales multiple, Exasol is trading at a premium to the group average on last year’s revenues, but looking at the higher growth peers, the valuation looks more in line.

The company’s financial performance and hence its share price will be sensitive to the following factors: