Darktrace specialises in responding to constantly evolving external threats by identifying common patterns. The company also makes cybersecurity software.
This week, the external threat came from New York-based Quintessential Capital Management, whose name is added to a long list of hedge funds and activist short sellers that have challenged Darktrace’s accounting practices. As the FT reported on Wednesday:
Quintessential’s allegations include: Darktrace appears to have simulated or anticipated sales to “phantom” customers through a “network of willing resellers”; that it seems to have incorrectly booked sales of hardware as software; and may have misrepresented the nature of its revenue.
Poppy Gustafsson, Darktrace chief executive, described Quintessential’s allegations as “unfounded inferences”. “I stand by my team and the business I represent,” she said in a statement on Wednesday, arguing the company has robust accounting and auditing practices.
Investors appear uncertain:
As usual, Darktrace’s defence rests largely on an argument that any weaknesses were addressed ahead of its flotation in September 2021. The Mike Lynch-backed company never goes as far as to apologise for its past life, choosing instead to talk of legacy issues born of corporate immaturity. Short sellers who see in each weakness an echo of Autonomy, Darktrace’s not-quite-parent company, are rarely appeased.
Most of Quintessential’s report is about revenue recognition. In response, Darktrace said it reviewed reseller controls during the IPO process and found a small number of contracts that did not come up to scratch, so they were excluded from the float.
Nearly all issues flagged are suggested to predate the 2020 appointment of CFO Cathy Graham, who was hired to bring “public market discipline” to the group. The allegations of incorrect booking of hardware sales, another Autonomy echo, are linked back to whether revenue should be recognised upfront or amortised over the kit’s useful life. Darktrace’s response points to its adoption of IFRS contracting accounting standards, again shortly before flotation.
Sellside analysts choose not to dwell on the past. “We believe that the group now has rigorous controls over revenue generated through partners,” says Numis, deeming Quintessential’s report “a swing and a miss”:
Our assessment is that QCM has likely identified an area of historic pre-IPO controls weakness in relation to Darktrace’s partner channel in some countries. However, this looks to be a weakness with limited scope, and which the group had addressed in 2020/21, before the IPO. We believe that the group now has rigorous controls over revenue generated through partners.
The remaining “new” issues raised by QCM . . . we think present no genuine concerns. We think that Darktrace’s financial statements are robust.
We expect QCM and others to continue to push a bear case, and many non-holders have pre-conceived ideas that these views play to. However, on this occasion we think that QCM has not meaningfully added to the debate.
Numis repeated its “buy” advice on the stock. Of the 11 brokers that have published on Darktrace this year, nine retain buy ratings and only one, Stifel, has downgraded.
Among the most bullish is Jefferies, whose price target of 425p is approximately double the current share price. Jefferies is Darktrace’s joint house broker, alongside Berenberg.
It told clients:
We have seen our fair share of debates in the sector, including names such as AIT, iSOFT, and Wirecard. While circumstances always vary, one consistent theme tends to be a mismatch between profits and cash flow. With that in mind, we think it is worth dwelling on the track record at Darktrace. Notably, cumulative cash flow (operating cash flow after capex and leases) from 2018-2022 of $58mn far outstrips the cumulative -$31mn of reported adj EBIT
Jefferies led Darktrace’s IPO, with Berenberg a global co-ordinator. Berenberg’s equity syndicate desk also handles much of the business for Darktrace early investors, including three secondary placings last year that raised a total of £585mn.
Another overlap is that Laura Janssens, Berenberg’s head of European equities since July 2020 and its former head of European research, is married to Darktrace head of investor relations Luk Janssens. Berenberg’s research on Darktrace has never noted the relationship.
Asked to comment, Berenberg said: “Our research department has systems and controls in place to manage and, where necessary, disclose any relevant conflicts of interest.” Laura Janssens added by email that she was “surprised and disappointed by the obvious inference of your questions”.
Berenberg’s research on Darktrace exceeds even Jefferies for bullishness, with analyst Benjamin May’s current share price target of 600p the second-highest in the City. He has maintained buy advice on the stock since starting coverage in June 2021 (except when restricted) and according to Bloomberg data has reiterated the call more than 30 times, most recently in a note published on Wednesday.
“Fundamentally, we believe that many of the [Quintessential] report’s conclusions are based on issues that have been taken out of context and are at minimum hard to validate,” Berenberg told clients.
Growing pains at Darktrace were first revealed in its 2019 results, released a year before the IPO, when it changed its reporting currency and adopted IFRS standards.* Spending on R&D was revised lower while employee headcount moved from research to sales, as highlighted by November 2021 report from research house Shadowfall, which has disclosed a short position:
These and similar discrepancies might be excused as start-up hype. Another example is on page 72 of the flotation prospectus, where a corporate funding history chart shows values were substantially below the amounts announced by Darktrace at the time. Series A is worth $10mn in the prospectus versus $18mn in the press release; Series C raised $34mn, not the $64mn reported; Series D is $50mn rather than the approximately $75mn that was announced; etc.
Darktrace said the mismatch was because the press releases rolled together the value each funding round’s of primary and secondary offerings, whereas the prospectus gave only the primary value.
Markedly more serious are the allegations of so-called channel stuffing — where resellers are compelled to buy product before they agree to an onward sale. Examples of potential channel stuffing in Quintessential’s report are “at best suspicious, and at worst fraudulent,” say Stifel analysts.
Berenberg argues otherwise. All examples appear to be linked to one individual who no longer works at Darktrace, it says, and “even if true we suspect they are immaterial to revenues”:
In short, we doubt this is a systemic issue across all of Darktrace’s operations. If it were, we suspect the report would have tried to present similar examples in Darktrace’s core markets (eg the UK, North America). After speaking with some of the UK’s largest listed VARs [value-added resellers] in recent days, some of whom are large partners of Darktrace, we came away encouraged by their comments that Darktrace’s contracts have top-tier governance in place. In short, they have not experienced any form of “channel stuffing”. We therefore suspect that any instances of this practice, if it has occurred in a few markets, are isolated and not widespread. We also believe that all allegations precede the company’s current CFO, whom we believe is highly prudent on these matters.
Quintessential’s report tackles the post-IPO period in part by drawing attention to non-current deferred revenues, which in this context means services paid for more than 12 months in advance. The ratio fell from around 33 per cent of sales in 2018 to just 9 per cent in 2022.
Might the trend be an indication that revenue has been inflated with unearned sales? Probably not, Berenberg says. Non-current deferred revenue has been shrinking because in recent years prepay “has become increasingly unusual” across the software industry, it told clients. And though Darktrace might have pushed for prepayments to generate positive cash flow when privately owned, such actions “would be unnecessary, given the company’s very significant cash balance and strong cash flow profile now.”
As part of its response, Darktrace on Wednesday launched a buyback of up to 35mn shares with a value of approximately £75mn, or about 5 per cent of its existing share capital. At the same time it announced a buyback for its employee benefit trust of to 10mn shares, at a net cost of £25mn, to cover future options.
Just 2.5mn Darktrace shares change hands through the London Stock Exchange on the average day, so the buybacks should help soak up most of the selling pressure. The combined amount exceeds total volume of Darktrace on the London Stock Exchange so far this week, which was around four times higher than average in the wake of the Quintessential report.
But absent a new defence strategy, will the company find any lasting relief? Not according to the Stifel analysts, who say “quirkiness” will keep making Darktrace “an easy target for shorts”:
“[W]e believe the company has failed since its IPO to alleviate concerns emanating from the market. While we believed time and strong operational results would help heal the company’s wounds, we are now of the view that the company, its Board and its management need to take more proactive and tangible measures. Among others, these could potentially include: 1) replacing part of the management and middle-management teams, 2) electing more independent Board members, 3) electing a new Chairman of the Board (current Chairman Gordon Hurst, while independent, is also Chairman of the Board of another company backed by Mike Lynch’s VC fund), 4) implementing an auditor rotation, and 5) launching a forensic audit run by another audit firm. Until material measures are taken, we believe more unpredictable hits are likely to affect the company’s share price in the future.
* CORRECTION: an earlier version of this post contained incorrect data on Darktrace’s revision to reported 2018 and 2019 sales. The data cited related to R&D spending, not total group revenue growth, which following the revisions were broadly unchanged.