How to get better at pitching fund selectors
Plus: lawyers discover GenAI and retail investors discover private equity
If you’ve been forwarded this email, you can sign up to the newsletter here.
What are you supposed to do in May, before you go away?
I had always thought it was "Buy" (as stocks waft upwards on summer optimism), but this year I learned the traditional phrase says it's "Sell in May and go away". Statistically speaking, you're better off buying, but — more pertinent to this newsletter — everyone in fund land seems to go away.
This being a quiet time in the financial-news cycle, I thought this week we could look at stories that offer a bunch of insight yet somehow didn't get all that much attention.
First up, a scientific experiment taken up by a law firm on how much work Generative AI (GenAI) can save. The law firm in question is Ashurst, and their report feels like they've just realised the scientific method is the best thing since sliced bread:
"From November 2023 to March 2024, Ashurst's Office of the Chief Digital Officer (the CDO team) led three global GenAI trials involving 411 partners, lawyers and staff representing all of our practice areas and business services functions across 23 offices in 14 countries in our global network. Through the trials, we sought to prove or disprove a series of working hypotheses concerning whether, how and how much our people would engage with and find value in GenAI."
They segmented their main findings under four categories.
1. VALUE:
Ashurst found the initial value of GenAI is in helping lawyers create first drafts quicker and more efficiently. Notice that just by saying "first draft" what Ashurst has discovered is that using GenAI involved creating a workflow.
If before you had said: "Junior lawyer, scuttle off and create this document" you're now saying: "Junior, use your AI to create a first draft and then review it." And the time these processes take varies considerably:
"In our controlled experiments, we measured approximate time savings of 80% to draft UK corporate filings requiring review and extraction of information from company articles of association, 59% to draft industry/sector-specific research reports that require reviewing and extracting key information from public company filings (Form 10-Ks), and 45% on creating first draft legal briefings." (Highlights mine.)
2. ACCURACY:
Ashurst reports GenAI-generated output can be difficult to distinguish from human output when legally correct. Let’s unpack. First, the 'legally correct' bit:
When they judged outputs just on accuracy, GenAI accuracy scores ranged from 1 to 4 out of 5, versus 3 to 4 out of 5 for humans. Either way you need to review the outputs, but with GenAI, you have to be extra vigilant.
It's a tricky deal; unlike rote work, it's hard for a humans to be vigilant to errors over longer terms, but it would seem a GenAI workflow demands this. Now let's turn to the part about 'difficult to distinguish':
"Interestingly, the expert panel correctly identified all outputs generated by our lawyers as human-generated. However, 50% of the output generated by GenAI were either misidentified as human-produced output or our experts couldn't tell whether the output was human- or AI-generated output...No human-generated output was misidentified as having been created by GenAI."
Human writing was entirely distinguishable from GenAI output, but in the other direction, you could not statistically tell!
Ashurst — by now thinking of ditching law altogether and turning to STEM — wanted to know what qualities made writing feel human. They feel it's a mix of language choice, tone and structure.
The term Ashurst was probably looking for is a sociolect, which is a sub-language that signals belonging to a specific social community. When AI models are trained, they are trained on huge language corpora, which "flatten out" the output language, giving it that tinny feel. There are also big issues to do with structure. And so arises the question of...
3. QUALITY:
Ashurst tried to find some quantitative measure for quality and quickly realised that's not going to happen:
"Quality in a legal context is multidimensional, with subjective and objective elements."
But it almost didn't matter. Even when the quality was rubbish, participants still felt the tools were valuable. Why?
4. SCOPE:
The above ties to the emotional value of working with GenAI tools. Ashurst employees reported that, beyond legal and client work, GenAI can make day-to-day life easier and...
"Tellingly, 88% of respondents at the end of the biggest trial said that using GenAI technology helped them to feel more prepared for the future."
I pay close attention to what financial services companies are doing with AI. So far, I've not seen anyone take it as seriously as Ashurst did here. This story got a little coverage in the business press, but most investment companies would do well to read this report in depth.
This newsletter takes 2 hours a week. Less work with more AI is possible.
We help marketing teams boost their output, through training and bespoke tools. Learn More.
Treasure Corner: Pitching to fund selectors
Last week the newsletter didn't go out because I was working hard on a PowerPoint deck. The short of it is that FinText got into a programme for female founders held by JP Morgan and Newable, that's centered on perfecting a pitch.
I was chuffed to be admitted as one of only 30 founders, and delighted to be among the top 8 selected for the final, which is next week. I'm telling you all this, here at Treasure Corner, for two reasons:
One, when we talk about marketing funds, we cover the journey right up to the point the portfolio manager enters the room to talk to a fund selector. But that meeting itself is a crucial aspect of selling, right? It's yet another form of marketing.
Years ago I interviewed a former fund selector on how he experiences those meetings, and the headline message was that PMs need to improve their pitching skills. Which brings me to the second reason I shared my little story. While building the deck, I stumbled upon analysis of several thousand startup pitch-decks to discover the most common pitfalls.
In a TechCrunch article titled "Here's where founders screw up their pitch decks most often", are the main findings, and most of them had to do with storytelling. Many seemed eerily relevant to selling funds.
For example, the study finds "Some 60% of founders didn’t have a clear enough target customer...". This matches up to when fund selectors say they want to understand how a fund meets a specific need in their portfolio. What kind of gap does it plug?
Here's another solid piece of advice:
"Just over half (55%) of the decks analyzed failed at the very first or last hurdle: Not having a strong introduction or closing slide. Last impressions matter, and a great summary and opening slide can help set the stage for the whole conversation."
Financial decks can be laden with charts and text, without any thought to what, if any, leaves an imprint in your customer's mind.
Finally, a piece of advice relevant to when portfolio managers articulate their strategies in the context of past performance — whether it was good or bad:
"68% of the decks didn’t anchor the company in time with a “why now” part of the story...Somewhere in your story it should become clear that the time is right: Why 10 years ago was too soon and why 10 years from now will be too late."
Also happening
Private assets have been all the rage this year, and summer calm has led the financial press to ask more questions. For example, the Wall St. Journal has been taking great interest in what happens when publicly visible funds interface with private equity assets.
First, it was about how some funds book superb one-day windfalls when buying private equity stakes on the secondary market. Instant markups lead to instant gains, all within the acceptable bounds of accounting rules, so funds' performance instantly improves, which helps attract more investment.
"Generally speaking, investors holding stakes in private-equity and other funds are allowed to estimate their fair value by relying on what the funds’ managers say they are worth, known as their net asset value (NAV)."
The accounting standards say investors using the NAV category may need to adjust this number. When these adjustments come, they're rarely on the upside. A couple of weeks ago, for example, BlackRock cut its $22 billion valuation of Byju, a former Indian edtech darling, to zero. BlackRock actually owned less than 1% of the company, but that's still a close to $200 million write-off.
The WSJ also looked at the pension funds that invested in PE and now struggle to exit.
"In the US, large public pension funds have an average 14% of their assets in private equity, while large corporate pensions have almost 13% in private equity and other illiquid assets such as private loans and infrastructure, according to data from Boston College and JPMorgan Chase."
Because the climate for exits has changed as interest rates rose, the PE funds struggle to distribute back the cash. It pushes them to sell assets on the secondary market.
"Secondary-market buyers last year paid an average of 85% of the value the assets were assigned three to six months before the sale, according to Jefferies Financial Group."
That's where the excellent single-day mark-ups are coming from! But pension funds are left strapped for cash.
All this isn't deterring retail investors who want in on private equity, so Blackstone is facing a luxurious dilemma: it's done a great job gathering assets, but in a market with few deals, it now looks to partner with smaller firms to invest it. Naturally, competitors like KKR and Carlyle are also pivoting to attract individual investors to their funds.
It's a shift in how high finance interacts with the masses. That is, they are now interacting, and I'm not sure the masses are better off for it.