Shares for the future: seven stocks removed from my top 40
Analyst Richard Beddard explains how he’s now working smarter rather than harder, and why he’s just kicked out so many companies from his Decision Engine.
20th June 2025 15:00
by Richard Beddard from interactive investor
For a decade or more, I’ve scored one company a week. I think of this as “feeding the machine”, but as I re-evaluate the machine (aka my Decision Engine) I’m questioning this approach.
- Our Services: SIPP Account | Stocks & Shares ISA | See all Investment Accounts
Decision Engine first
Two things have changed. I’m much more familiar with longstanding Decision Engine shares. And the world has become a less familiar and less predictable place.
Taking one company at a time and spreading my effort evenly, let us call it a “company first” approach, no longer feels optimal.
Businesses are facing a greater range of challenges, sometimes requiring me to go deeper to establish whether I can be confident enough to invest.
And these challenges have exposed weaknesses, diminishing some firms’ appeal and requiring me to cast my net wider to find new ideas.
The times are directing me to go deeper and wider, but such a strategy is incoherent with finite resources. You roll the pin over the pastry backwards and forwards and left and right until it rips in the middle. It is a recipe for burnout, or a bad pie.
The answer is to work smarter rather than harder. There are two ways I’m trying to do this: the first is artificial intelligence (AI). The second is a Decision Engine first approach.
So far, my experience of AI has made me less efficient. It has increased my workload because of the number of rabbit holes you can burrow down when you can ask anything, and the amount of fact-checking fabulating AI generates.
But AI has improved, and I’ve become more disciplined. I believe I’m on the cusp of break-even now, when it comes to efficiency.
To be clear, I use AI to query annual reports and other documents for information and I check what I learn against the original documents.
- Stockwatch: choosing between these two mid-cap shares
- Why Warren Buffett could back Britain in final mega-deal
I think by writing my arguments. Then I read them aloud to make sure they cohere and flow. These become my articles.
Maybe one day AI will retire me, but until then, it will be useful. Nothing you read has been or will be written and consequently “thought” by AI.
A “Decision Engine first” approach means each week, when I plan this article, I will not be asking: “Which company will I score?” as I used to. I will be asking: “What do I need to know to improve the Decision Engine?”
Sometimes this will require me to score a share. For example, when I add one to the Decision Engine for the first time.
Sometimes it may only require me to update a score. But sometimes I will need to go deep to explore the significance of an emerging risk.
The innovation that has given me the flexibility to try the “Decision Engine first” approach is the new summary score format. This is generated by the Decision Engine spreadsheet and is easily appended to an article.
To illustrate, let us take a look at Celebrus Technologies (LSE:CLBS). Recent events mean an update is overdue.
Celebrus: material risk materialises
Celebrus is a former systems integrator that acquired a marketing data software platform in 2015. Since then, it has focused on the software-side of the business, and today it only supplies third-party hardware and software to a diminishing band of existing customers. I last scored it in August 2024.
Celebrus achieved a speculative score (7 or less), and in April I liquidated the Share Sleuth portfolio’s holding just before it issued a trading update that said revenue in the year to March 2025 was lower than it was in 2024 (although profit was higher). The shares were pricey according to the Decision Engine, and I didn’t feel sufficiently confident in my understanding of the business or the market to hold them.
Celebrus’ journey has involved wholesale changes to its business model, and indeed its name, which was previously D4t4. To recap: systems integrators supply IT systems sourced from a number of vendors. In becoming one of those vendors its business model is changing radically.
For a start, like many software houses, the company has converted Celebrus into a cloud-based system that works out of the box, changing how it charges for the service and reducing the level of support required. It has also developed a product variant that detects fraud.
Celebrus’ sales strategy has changed. Mostly the company still relies on systems integrator partners to sell the product, but it has established a direct sales team to create end-user demand and reduce this dependency. This is enabling it to sell to a broader range of customers than its niche of banks and financial institutions.
- AIM’s 30th birthday: best stocks and what the future holds
- Sign up to our free newsletter for investment ideas, latest news and award-winning analysis
Celebrus touts compelling features, elements of which are patented. It works in real time to shape website and app users’ experiences using biometrics and machine learning, and it relies on first party (customer’s own data). This makes it easier to comply with privacy regulations.
The big risk, I felt, was Celebrus’ reliance on a systems integrator partner for nearly two-thirds of revenue that year, and substantial revenues in most years.
There was also an issue with the quality of revenue, which includes volatile hardware sales to legacy customers still installing software on-site and third-party software sales.
To my mind, Celebrus’ trading update was written in the wrong order. The least significant piece of information was the first, that revenue would decline as customers delayed investment due to the scary geopolitical context.
More important was the next bit. Celebrus has redefined its definition of Annual Recurring Revenue (ARR) to incorporate a change in accounting policy. From 2025, it will recognise revenue on a monthly basis, recognising it more gradually than the old annual policy, depressing revenue growth.
More significantly, to my mind, ARR will exclude third-party software sales and focus on Celebrus’ licenses and hosting, the company’s own intellectual property and strategic focus.
Celebrus is the future of the business and, I believe, its most profitable component. The company says ARR has increased 13.9% under the new methodology. This is good news, but the company hasn’t told us how much revenue Celebrus earns, how profitable it is, and how much it has grown in previous years.
We got some of this information in 2022, when Celebrus attributed £6.1 million revenue for the Celebrus family of products, less than half of product revenue and about a quarter of total revenue, including various services.
The board is promising enhanced clarity and disclosure in the full-year results to be published next month. At least until then, though, the need for clarity and disclosure only serves to highlight the confusion that led me to lose confidence in the business.
Another piece of significant news was tagged on the end of the trading update, and may have precipitated the change in accounting policy. Celebrus is losing a chunk of partner-led third-party software sales that were part of the old definition of ARR.
In other words, the revenues haven’t recurred. Part of the partner risk has materialised, and non-Celebrus revenue is likely to be lower in 2026.
The combined impact of lost hardware sales and the new accounting policy on the company’s revenue and profit in 2026, was still to be determined at the time of the trading update, although we should learn more about this in the full-year results.
The conflation of legacy revenues with Celebrus revenues and changes in accounting policy mean I do not trust the historical numbers. Maybe the full-year results will satisfy me, but I wonder if management should have been more forthcoming sooner.
Celebrus’ falling share price lifted the company into value territory. Just before the re-evaluation below, it scored 7.8. Until I have more data though, I am not that confident:
Celebrus | CLBS | Marketing and fraud protection software as a service | 18/06/2025 | 6.3/10 |
How capably has Celebrus made money? | 2.0 | |||
Celebrus has remained profitable and cash generative as it has changed business model from systems integration to the supply of software with some impressive features. Its track record is difficult to interpret because of volatile revenues from legacy hardware and software sales. | ||||
How big are the risks? | 2.0 | |||
Although Celebrus has established a direct sales force it is still highly dependent on resellers and integrations with enterprise software vendors. It competes with a wide range of software providers with overlapping capabilities that I cannot differentiate | ||||
How fair and coherent is its strategy? | 2.0 | |||
Celebrus is backing its own patented IP, by selling it more aggressively and selling into new markets. The strategy may be coherent but without better data I can't verify that it is winning. Management is promising clarity, but it could have been more forthcoming in the past. | ||||
How low (high) is the share price compared to normalised profit? | 0.3 | |||
Low. A share price of 152p values the enterprise at £31 million, about 17 times normalised profit. | ||||
A score of 6.3/10 indicates Celebrus is a somewhat speculative investment. | ||||
NB: Bold text indicates factors that reduce the score. Bold and italicised text doubly so. The maximum score is 3 for each criterion except price, which has a maximum of 1 (explainedhere) | ||||
22 Shares for the future
Here is the ranked list of Decision Engine shares. I review the scores at least once a year, soon after each company has published its annual report. The price scores are calculated using the share price prior to publication.
Generally, I consider shares that score more than 7 out of 10 to be good value. Shares that score 7 or less are good businesses but they are not obviously cheap at the moment.
Celebrus’ score for quality, the combination of its scores for Capabilities, Risks and Strategy, has dropped to 6. Since the maximum score for price is 1, this leaves it with a maximum potential score of 7 depending on the share price.
Until I reappraise the share it cannot, therefore, be good value. I have removed Celebrus along with five other stranded shares from the Decision Engine. They are: Marks Electrical Group Ordinary Shares (LSE:MRK), RWS Holdings (LSE:RWS), Tracsis (LSE:TRCS), Treatt (LSE:TET) and Victrex (LSE:VCT).
I will not forget about these shares. They are now in my research list and must duke it out with new candidates if they are to re-enter the table.
I have also removed Garmin Ltd (NYSE:GRMN) because I simply don’t know how to score an excellent business that has almost its entire manufacturing footprint in Taiwan, a geopolitical hotspot.
The removal of these shares refocuses the Decision Engine on shares I’d consider for investment at the right price. Consequently, I have reunited the shares that are good value and the ones that might be, should they become cheaper.
I’ve also added columns for quality and price, so you can see how the score has been determined. For example, top-scoring company Latham (James) (LSE:LTHM) scores 8 for quality and 1 for price, giving a total of 9. Bottom-scoring Keystone Law Group Ordinary Shares (LSE:KEYS) scores 7 for quality and -2.4 for price, giving a total of 4.6.
Company | Description | Score | Qual | Price | |
1 | James Latham | Imports and distributes timber and timber products | 8.0 | 1.0 | |
2 | FW Thorpe | Makes light fittings for commercial and public buildings, roads, and tunnels | 8.5 | 0.3 | |
3 | Dewhurst | Manufactures pushbuttons and other components for lifts and ATMs | 7.5 | 1.0 | |
4 | Churchill China | Manufactures tableware for restaurants etc. | 7.5 | 1.0 | |
5 | Howden Joinery | Supplies kitchens to small builders | 8.0 | 0.4 | |
6 | Oxford Instruments | Manufacturer of scientific equipment for industry and academia | 7.5 | 0.8 | |
7 | Renishaw | Whiz-bang manufacturer of automated machine tools and robots | 7.5 | 0.6 | |
8 | Focusrite | Designs recording equipment, loudspeakers, and instruments for musicians | 7.0 | 1.0 | |
9 | Solid State | Assembles electronic systems (e.g. computers and radios) and distributes components | 7.5 | 0.5 | |
10 | Bunzl | Distributes essential everyday items consumed by organisations | 7.5 | 0.5 | |
11 | Hollywood Bowl | Operates tenpin bowling and indoor crazy golf centres | 7.5 | 0.3 | |
12 | Macfarlane | Distributes and manufactures protective packaging | 7.0 | 0.7 | |
13 | Renew | Repair and maintenance of rail, road, water, nuclear infrastructure | 7.5 | 0.2 | |
14 | YouGov | Surveys and distributes public opinion online | 7.5 | 0.1 | |
15 | Anpario | Manufactures natural animal feed additives | 7.0 | 0.6 | |
16 | James Halstead | Manufactures vinyl flooring for commercial and public spaces | 7.0 | 0.5 | |
17 | Advanced Medical Solutions | Manufactures surgical adhesives, sutures, fixation devices and dressings | 7.0 | 0.5 | |
18 | Games Workshop | Manufactures/retails Warhammer models, licences stories/characters | 9.0 | -1.7 | |
19 | Jet2 | Flies holidaymakers to Europe, sells package holidays | 7.5 | -0.3 | |
20 | Porvair | Manufactures filters and laboratory equipment | 8.0 | -0.8 | |
21 | Bloomsbury Publishing | Publishes books, and digital collections for academics and professionals | 7.5 | -0.3 | |
22 | Auto Trader | Online marketplace for motor vehicles | 8.0 | -0.9 | |
23 | Softcat | Sells hardware and software to businesses and the public sector | 8.0 | -1.2 | |
24 | Volution | Manufacturer of ventilation products | 8.0 | -1.2 | |
25 | 4Imprint | Customises and distributes promotional goods | 8.0 | -1.2 | |
26 | Dunelm | Retailer of furniture and homewares | 8.0 | -1.4 | |
27 | DotDigital | Provides automated marketing software as a service | 6.5 | -0.1 | |
28 | Quartix | Supplies vehicle tracking systems to small fleets and insurers | 7.5 | -1.3 | |
29 | Judges Scientific | Manufactures scientific instruments | 7.5 | -1.5 | |
30 | Goodwin | Casts and machines steel. Processes minerals for casting jewellery, tyres | 8.0 | -2.3 | |
31 | Tristel | Manufactures disinfectants for simple medical instruments and surfaces | 7.5 | -2.1 | |
32 | Cohort | Manufactures military technology, does research and consultancy | 7.5 | -2.3 | |
33 | Keystone Law | Runs a network of self-employed lawyers | 7.0 | -2.4 |
Scores and stats: Richard Beddard. Data: ShareScope and annual reports.
Click on a share's score to see a breakdown (scores may have changed due to movements in share price).
Richard Beddard is a freelance contributor and not a direct employee of interactive investor.
Richard owns many shares in the Decision Engine. He weights his portfolio so it owns bigger holdings in the higher-scoring shares.
For more on the Decision Engine, please see Richard’s explainer.
Contact Richard Beddard by email: richard@beddard.net or on Twitter: @RichardBeddard
AIM stocks tend to be volatile high-risk/high-reward investments and are intended for people with an appropriate degree of equity trading knowledge and experience.
These articles are provided for information purposes only. Occasionally, an opinion about whether to buy or sell a specific investment may be provided by third parties. The content is not intended to be a personal recommendation to buy or sell any financial instrument or product, or to adopt any investment strategy as it is not provided based on an assessment of your investing knowledge and experience, your financial situation or your investment objectives. The value of your investments, and the income derived from them, may go down as well as up. You may not get back all the money that you invest. The investments referred to in this article may not be suitable for all investors, and if in doubt, an investor should seek advice from a qualified investment adviser.
Full performance can be found on the company or index summary page on the interactive investor website. Simply click on the company's or index name highlighted in the article.
Disclosure
We use a combination of fundamental and technical analysis in forming our view as to the valuation and prospects of an investment. Where relevant we have set out those particular matters we think are important in the above article, but further detail can be found here.
Please note that our article on this investment should not be considered to be a regular publication.
Details of all recommendations issued by ii during the previous 12-month period can be found here.
ii adheres to a strict code of conduct. Contributors may hold shares or have other interests in companies included in these portfolios, which could create a conflict of interests. Contributors intending to write about any financial instruments in which they have an interest are required to disclose such interest to ii and in the article itself. ii will at all times consider whether such interest impairs the objectivity of the recommendation.
In addition, individuals involved in the production of investment articles are subject to a personal account dealing restriction, which prevents them from placing a transaction in the specified instrument(s) for a period before and for five working days after such publication. This is to avoid personal interests conflicting with the interests of the recipients of those investment articles.