Tungsten Network: e-invoicing activities – targeting profit
e-invoicing is a large and growing business. Billentis, an e-invoicing consultancy firm, estimates that globally, around 170bn invoices are sent by/to businesses and governments each year and of these around 28bn or 16%, are e-invoices and this figure is rising. Currently there is no official organisation collating the statistics in a standardised way.
Billentis estimates that e-invoices can save 60-80% of the cost of a paper invoice and quotes a Finnish state treasury estimate of a paper invoice costing €30-50 to process versus €1 for a fully automated one, and an Italian study by Politecnico di Milano quoting savings of €65 per invoice for a fully integrated process.
According to data collected by Billentis, most European e-invoicers just about break even in their
e-invoicing activities, as most of the cost savings made by using e-invoicing accrue to users rather than system providers such as Tungsten Network. Billentis also estimates that over 70% of the companies offering e-invoicing in Europe were part of companies that were profitable overall thanks to their other activities, which allow them to finance expansion of their e-invoicing activities. Billentis believes that e-invoicers seek market share and scale and hope to eventually make profits from providing add-on products such as spend analytics and invoice financing. Two leading e-invoicers, Ariba and Basware also offer these add-on services.
Tungsten’s management has an ambitious plan to achieve profitability through repricing and adding many more suppliers and buyers. Tungsten offers invoice financing which, following the sale of the bank, will be financed via the arrangement with Insight. Other funding sources are being pursued. In September 2013, Tungsten acquired the rights via a licence arrangement (terms not disclosed) to a spend analytics business. Tungsten Network has a trading history of over 10 years and around £50m has been invested in the development of the network. The business has never made a profit and, as we show in the table below, the EBITDA/revenue ratio has been negative by 10% to 18% over the last few years.
Exhibit 1: Tungsten Network*
£m |
2011 |
2012 |
2013 |
2014 |
2015* |
H116 |
Revenue |
15.327 |
16.34 |
17.631 |
19.501 |
23.018 |
12.976 |
EBITDA |
(2.372) |
(2.432) |
(2.564) |
(2.747) |
(2.296) |
(2.292) |
EBITDA/revenue (%) |
(15) |
(15) |
(15) |
(14) |
(10) |
(18) |
Source: Tungsten. Note: *Underlying EBITDA after excluding one-off costs.
At end H116, Tungsten had 178 buyers on its network and 194,500 suppliers. The former earned 39% of revenues in that period and the latter 61%. Supplier revenue amounted to around £84 each in H116 and included annual licence and transactions fees. Only around 20% of suppliers pay fees to Tungsten, as Tungsten does not charge them for the first 52 invoices that are processed. Revenue from buyers was around £58,000 each and was mainly from licence/transaction fees but also included some set-up fees. The buyers on the network are mainly large, often well-known multinationals or large government agencies. On average, each buyer has over 1,000 suppliers who are frequently SMEs. Tungsten operates a closed network, which requires that buyers on its network invite their suppliers to join them on the network.
Exhibit 2 shows what needs to be achieved by e-invoicing according to management guidance for the group to break even in FY17.
Exhibit 2: Tungsten Network
Tungsten Network |
2015 |
H116 |
2016e |
2017e |
2018e |
2019e |
2020e |
Suppliers – end period |
181,000 |
194,500 |
208,000 |
236,000 |
266,000 |
298,000 |
333,000 |
Increase in suppliers |
13,000 |
13,500 |
27,000 |
28,000 |
30,000 |
32,000 |
35,000 |
Average suppliers |
174,500 |
187,750 |
194,500 |
222,000 |
251,000 |
282,000 |
315,500 |
Revenue per supplier (£) |
75 |
84 |
75 |
75 |
75 |
75 |
75 |
Increase in revenue per supplier |
|
|
0% |
0% |
0% |
0% |
0% |
Growth in average suppliers |
13% |
15% |
11% |
14% |
13% |
12% |
12% |
Buyers – end period |
173 |
178 |
187 |
202 |
217 |
232 |
247 |
Increase in buyers |
49 |
6 |
15 |
15 |
15 |
15 |
15 |
Average buyers |
156 |
176 |
180 |
196 |
211 |
226 |
241 |
Increase in average buyers |
|
|
15% |
9% |
8% |
7% |
7% |
Revenue per buyer (£) |
63,583 |
58,000 |
71,271 |
86,238 |
108,660 |
108,660 |
108,660 |
Increase in revenue per buyer |
|
194,500 |
12% |
21% |
26% |
0% |
0% |
Revenue |
|
|
|
|
|
|
|
Suppliers |
13.1 |
7.9 |
14.6 |
16.7 |
18.9 |
21.2 |
23.7 |
Buyers |
9.9 |
5.1 |
12.8 |
16.9 |
22.9 |
24.5 |
26.1 |
Total |
23.0 |
13.0 |
27.4 |
33.5 |
41.7 |
45.7 |
49.8 |
Administrative expenses |
(28.2) |
(15.7) |
(31.7) |
(32.0) |
(32.0) |
(32.0) |
(32.0) |
EBITDA |
(5.1) |
(2.7) |
(4.3) |
1.5 |
9.7 |
13.7 |
17.8 |
One-off costs |
2.8 |
0.4 |
1.0 |
|
|
|
|
Underlying EBITDA |
(2.3) |
(2.3) |
(3.2) |
1.5 |
9.7 |
13.7 |
17.8 |
Underlying EBITDA margin |
(10%) |
(18%) |
(13%) |
5% |
23% |
30% |
36% |
Source: Edison Investment Research, Tungsten
In H116, Tungsten Network added another six buyers, having added a similar number in FY15 (excluding the 43 from the acquisition of DocuSphere) and two in FY14. This results in a total number of 178 buyers (two buyers merged in H116). Tungsten’s efforts to increase buyer numbers do seem to be bearing fruit; we have assumed nine more buyers in H216 and 15 per year in FY17 to FY20. Tungsten Network added 13,500 suppliers in H116, slightly ahead of the FY15 figure, but less than it planned. Tungsten has said that there are many suppliers in the pipeline waiting to join the network, but it is taking longer than expected to on-board them. Tungsten is taking steps to speed up the process, for instance by taking what was a mainly manual upload process carried out by Tungsten to a digital process performed over the internet and supported by the buyers who nominate the suppliers. We have assumed a similar level of suppliers is on-boarded in H216 and thereafter the number increases by a few thousand each year to reach 35,000 new suppliers added in FY20.
Raising prices for buyers
Tungsten intends to persuade its buyers that they should pay more for the service by convincing them of the following:
■
that on average each buyer processes just under 100,000 invoices per year and assuming that they save 50% of the cost savings calculated by the Finnish state treasury, the buyers would save £10 on each invoice processed making total savings of £1m per year, far in excess of the fee they are currently paying; and
■
Tungsten will offer its analytics service to buyers, which could enable them to make further cost savings. The analytics service analyses a buyer’s spending record and identifies occasions when it has paid one suppler more than another for the same good or service. As each buyer on average processes around £800m of invoices through Tungsten, if the analytics service enabled 10bp of saving it could enable each buyer to make further savings of around £0.8m.
As a buyer can take up to 18 months to connect its IT system to Tungsten Network’s and on-board its suppliers, there is likely to be considerable inertia and reluctance among the buyers to move to another e-invoicer, so Tungsten Network does have some pricing power. It now proposes to use that power. Tungsten Network’s contracts with buyers normally last three years, so that every year around one-third come up for renewal. In H116, Tungsten renegotiated renewals with 14 buyers at price increases averaging 70%. A further 35 buyers have contracts coming up for renewal in H216. In our forecasts we have assumed that Tungsten achieves a similar level of price renewal over the next two and a half years so that the average price increases are around 12% in FY16, 21% in FY17 and 26% in FY18. It should be noted that H116 revenue per buyer was slightly lower than it was in FY15 because the majority of new buyers were signed at the end of the period, so that little revenue was earned by them in the period and there were fewer, large one-off buyer set-up fees in H116 compared with the previous year. In our forecasts we have assumed that revenue per supplier remains around £75 in H216 and from FY17 to FY20. Tungsten expects supplier revenues to increase as a result of the Network effect – in particular, as new buyers are added Tungsten expects to earn additional revenue from connecting current suppliers to them.
Tungsten Network administrative expenses
In H116 administrative expenses were £15.7m, with just £0.4m of one-offs. Tungsten Network is proposing to keep administrative costs in H216 at a similar level to those in H116, making a total of £31.7m for the year. Thereafter it proposes to keep costs flat. Tungsten says that it has re-engineered its processes in Tungsten Network to reduce reliance on manual intervention and increase the amount of automation and tasks carried out by buyers and suppliers and this and similar measures will enable it to control costs. We have little visibility over the measures that have been taken and their likelihood of success so the cost control ambitions remain subject to considerable uncertainty.
Our forecasts anticipate that network revenue will increase 19% in FY16 and 22% in FY17, and generate a 14% CAGR from FY17-20. These calculations are at best tentative and rely heavily on Tungsten Network’s sales efforts being successful. We have assumed that administrative costs are £32m pa from 2017-20. On the basis that Tungsten does control its costs as it intends, we calculate that Tungsten Network may achieve an EBITDA of £1.5m in FY17, EBITDA of £9.7m in FY18 and an EBITDA margin of 36% in FY20.
Invoice financing: A possible upside
While Tungsten seeks to earn profit from e-invoicing, it aims to enhance its returns by cross-selling the invoice financing product to the suppliers on its network. Invoice financing enables suppliers to obtain cash earlier than they would do otherwise and so improve their cash flow. A supplier sells an invoice that they have raised to a buyer of their product or service at less than face value to an invoice financier, frequently a subsidiary of a well-established bank. The invoice financier should then receive full payment on the invoice a little later and earn the difference between the amount received and the amount paid. Tungsten seeks to use the supplier relationships that it has acquired from its e-invoicing activities to build an invoice financing business. It proposes to make its method of giving credit, called Tungsten Early Payment (TEP), superior to the incumbent operators. The delivery mechanism will be fully integrated with the e-invoicing activities and once set up, will require an authorised individual to click a box on a list of invoices available for early payment.
TEP intends to offer a low cost and transparent pricing structure, in contrast to many existing operators who currently have a complicated fee structure and charge relatively higher prices.
Initially funding was to be provided by Tungsten Bank. However, the lower than expected take-up of its invoice financing offer, together with the availability of finance from others has led Tungsten to conclude that shareholders’ interests would be better served by selling the bank. An agreement to sell it for £30m, NAV (ex-intangibles) plus c £4m-5m premium depending on when the transaction closes, was announced with the H116 results and the transaction is expected to be completed within six to 12 months.
In December 2014, Tungsten announced that it had entered into an agreement with Insight Investment Management, under which Insight agreed to acquire invoices from Tungsten. Insight is a global investment manager owned by BNY Mellon, with over £300bn of assets under management. In addition to Insight there are many other potential acquirers of discounted invoices, including alternative finance providers and some of the recently listed investment companies investing in SME loans. Details of Insight’s revenue sharing arrangements with Tungsten are not in the public domain. We expect that Insight’s funds will bear any credit loss from the lending and Tungsten will take a share of the interest received from the lending, which it will record in its accounts as a fee. The lending will not be recorded on Tungsten’s balance sheet (nor that of Tungsten Bank) so there will be no requirement for capital backing for the lending to comply with Bank of England PRA capital requirements. The anticipated lending rates depend on the size of the supplier, with SMEs estimated to pay up to 12% per year and large companies around 3% to 5% per year.
Tungsten launched its invoice financing service in November 2014. As evident in the H116 results, the take-up has been slower than hoped for. For a supplier to discount an invoice it is necessary that Invoice Status Service (ISS) is enabled for buyers. It is also necessary for suppliers to register for the service. At the end of H116, 294 suppliers had registered (with 108 using the service), up from 38 at end FY15, but this is a small fraction of the number of suppliers on the network. The value financed was £43m, from £32m at end FY15. Most of these invoices were of short duration, 40 days in H116, such that the average amount of outstanding finance in H116 was £12.8m. While this is a meaningful increase on the £6.5m in H215, it is a small fraction of the amount that its suppliers likely financed during the period, which we estimate was probably around £1bn. The average gross yield on the financing was 6.17%.
The potential to sell invoice financing to suppliers on its network remains, and indeed other e-invoicers are also offering suppliers invoice financing such as Basware and Ariba, a large e-invoicer owned by SAP. Tungsten faces competition for invoice financing not just from the traditional suppliers but also other e-invoicing organisations.
Tungsten has identified two main causes for the slow take-up of this product to date:
■
the procedure to enable suppliers to qualify for invoice financing has been too cumbersome and time consuming and has discouraged many from applying. Tungsten is seeking to remedy this by simplifying and speeding up the approval process; and
■
in many cases, the person deciding whether to obtain finance from discounting invoices is unfamiliar with Tungsten and is based in a different department, treasury rather than accounts receivable. Tungsten has invested in more sales and marketing to overcome these problems.
Additionally Tungsten needs to work hard to establish its brand and reputation as a provider of finance.
In H116 revenue from TEP was £84,000 and consisted of the revenue on invoice receivables financed by Tungsten Bank. No revenue was recognised on invoice receivables financed by Insight as Tungsten could not provide invoices that met with Insight’s requirements
Invoice financing forecasts
Our tentative forecasts for Tungsten Network Finance are shown in Exhibit 3 below. For H216 and H117 we have assumed a similar level of revenue and profits as H116, but for H217 and onwards we have assumed the sale of the bank has completed and all revenues will be earned by passing the invoices to Insight or another funding provider. We have assumed that Tungsten Network earns fees of c 1% of the value of these invoices. This fee is towards the lower end of fees earned by other providers. GLI Finance, a company that owns 19 alternative finance platforms arranging loans for SMEs, earns fees of c 1-4% for loans arranged. Given Tungsten’s desire to build this business, we have assumed, conservatively, that it charges fees at the bottom of this range. We have anticipated that invoices discounted in H216 are 15% higher than in H1 and that from FY17-20 they grow at a rate of 20% pa. In H116, an impairment charge of £6.3m was incurred as a result of the propose sale of the bank, which should be partially recovered in FY17 as a gain of £2.2m should be earned on completion of the proposed disposal. We have assumed that the cost base falls by £1m in FY17 as a result of the sale of the bank and by £2m per year thereafter, as indicated by management. These forecasts anticipate that EBITDA losses fall from £5.4m in FY17 to £2.5 in 2020. We have taken a cautious view of Tungsten Network Finance forecasts given the lack of visibility on its likely volume growth or its pricing structure. We expect that Tungsten will persist with the activity given the potential size of the market and the likely reduction in losses.
Exhibit 3: Tungsten Network Finance
£m |
2015 |
H116 |
2016e |
2017e |
2018e |
2019e |
2020e |
Invoices discounted £m |
|
65 |
140 |
168 |
202 |
242 |
290 |
Growth pa |
|
|
|
20% |
20% |
20% |
20% |
Average TEP lending |
3.3 |
12.8 |
14.8 |
25.0 |
30.0 |
40.0 |
50.0 |
Growth pa |
|
|
|
69% |
20% |
33% |
25% |
Revenue |
|
|
|
|
|
|
|
Fees |
0.1 |
0.1 |
0.3 |
1.0 |
2.0 |
2.4 |
2.9 |
Total revenue |
0.1 |
0.1 |
0.3 |
1.0 |
2.0 |
2.4 |
2.9 |
Administrative expenses |
(12.8) |
(3.7) |
(7.4) |
(6.4) |
(5.4) |
(5.4) |
(5.4) |
EBITDA |
(12.7) |
(3.6) |
(7.1) |
(5.4) |
(3.4) |
(3.0) |
(2.5) |
One-off costs |
6.3 |
1.3 |
2.8 |
0.0 |
|
|
|
Underlying EBITDA |
(6.4) |
(2.4) |
(4.4) |
(5.4) |
(3.4) |
(3.0) |
(2.5) |
Underlying EBITDA margin |
N/A |
N/A |
N/A |
N/A |
N/A |
N/A |
N/A |
Source: Edison Investment Research, Tungsten