Investment Manager Sheds $45.7 Million Worth of Software Stock, According to Latest SEC Filing

Conestoga Capital Advisors sold 634,534 shares of SPS Commerce in Q1 2026, an estimated $45.72 million transaction that cut the fund’s stake to 8,270 shares valued at about $460,391. The position’s share of AUM fell from roughly 1.0% to about 0.01%, and the quarter-end stake value dropped $56.83 million including price declines. While the sale signals reduced conviction, the company still showed resilient revenue growth and trades at a 2.8x price-to-sales multiple near decade lows.

Analysis

This looks less like a simple trim and more like a forced de-risking out of a crowded software winner into a name with deteriorating momentum and little institutional sponsorship support. When a long-duration compounder falls from a meaningful book weight to effectively de minimis, the signaling effect can matter more than the dollar amount: it often tells you the manager’s expected forward IRR has fallen below the hurdle relative to higher-conviction industrial and data names. The second-order effect is that the stock loses a natural source of buy-the-dip liquidity just as fundamental investors become more selective on software names with only mid-single-digit growth.

The main risk is that this becomes a self-reinforcing multiple compression event. A sub-3x sales multiple can look cheap, but for a recurring-revenue vertical software business it only becomes a true floor if growth stabilizes and net retention doesn’t roll over; otherwise valuation can still reset another 20-30% before it finds durable support. The setup is most vulnerable over the next 1-3 quarters if management commentary fails to translate AI product work into measurable bookings or if customers delay workflow upgrades amid broader IT budget scrutiny.

The contrarian case is not that the stock is obviously cheap, but that the market may be extrapolating cyclical weakness into a structural problem. If the company can maintain revenue growth while proving the AI initiative improves implementation speed or wallet share, the stock can re-rate quickly because expectations are already deeply washed out. In that sense, the better trade is not a blind outright long; it is a catalyst-driven position that only pays if fundamentals stop deteriorating faster than the multiple already implies.

Key Decisions for Investors

• Avoid catching the knife in SPSC outright; wait for a 1-2 quarter inflection in billings or commentary on retention before buying, because downside can persist another 20-30% if growth decelerates.
• For investors who want exposure to supply-chain software but lower idiosyncratic risk, favor a basket long in CWST/BCPC/FSV/RBC over SPSC: these names are benefiting from steadier institutional sponsorship and cleaner relative momentum.
• Consider a pairs trade: long a high-quality software name with stabilizing fundamentals, short SPSC for 3-6 months, targeting further multiple dispersion if the market continues rewarding quality over value traps.
• If you want optionality on a recovery, use a small-risk call spread in SPSC dated 6-9 months out; the trade only works if management proves AI can reaccelerate growth, but downside is capped and the stock is already priced for disappointment.
• Do not interpret the sale as a short signal by itself; treat it as a position-sizing and opportunity-cost signal, and reassess only if additional holders start cutting exposure in the next filing cycle.